Daily Archives: October 16, 2023

Xiao, Y., Zhai, Q., Zhang, H., Chen, W. & Hill, C. Gut colonization mechanisms of Lactobacillus and Bifidobacterium: an argument for personalized designs. Annu. Rev. Food Sci. Technol. 12, 213233 (2021).

Article CAS PubMed Google Scholar

Walter, J., Maldonado-Gomez, M. X. & Martinez, I. To engraft or not to engraft: an ecological framework for gut microbiome modulation with live microbes. Curr. Opin. Biotechnol. 49, 129139 (2018).

Article CAS PubMed Google Scholar

Mallon, C. A., Elsas, J. D. V. & Salles, J. F. Microbial invasions: the process, patterns, and mechanisms. Trends Microbiol. 23, 719729 (2015).

Article CAS PubMed Google Scholar

Porter, N. T. & Martens, E. C. Love thy neighbor: sharing and cooperativity in the gut microbiota. Cell Host Microbe 19, 745746 (2016).

Article CAS PubMed Google Scholar

Russell, B. J. et al. Intestinal transgene delivery with native E. coli chassis allows persistent physiological changes. Cell 185, 32633277 (2022).

Article CAS PubMed PubMed Central Google Scholar

Wang, J. et al. Genome adaptive evolution of Lactobacillus casei under long-term antibiotic selection pressures. BMC Genomics 18, 320 (2017).

Article PubMed PubMed Central Google Scholar

Huang, S. et al. Candidate probiotic Lactiplantibacillus plantarum HNU082 rapidly and convergently evolves within human, mice, and zebrafish gut but differentially influences the resident microbiome. Microbiome 9, 151 (2021).

Article CAS PubMed PubMed Central Google Scholar

Anal, A. K. & Singh, H. Recent advances in microencapsulation of probiotics for industrial applications and targeted delivery. Trends Food Sci. Technol. 18, 240251 (2007).

Article CAS Google Scholar

Gao, P. et al. Screening and identification of probiotic bifidobacterium from Mongolian children. Acta Microbiol. Sin. 49, 210216 (2009).

CAS Google Scholar

Sun, Z. et al. Complete genome sequence of probiotic Bifidobacterium animalis subsp. lactis strain V9. J. Bacteriol. 192, 40804081 (2010).

Article CAS PubMed PubMed Central Google Scholar

Yan, Y. et al. Probiotic Bifidobacterium lactis V9 attenuates hepatic steatosis and inflammation in rats with non-alcoholic fatty liver disease. AMB Express 10, 101 (2020).

Article CAS PubMed PubMed Central Google Scholar

Diaz, R., Torres-Miranda, A., Orellana, G. & Garrido, D. Comparative genomic analysis of novel Bifidobacterium longum subsp. longum strains reveals functional divergence in the human gut microbiota. Microorganisms 9, 1906 (2021).

Article CAS PubMed PubMed Central Google Scholar

Tenaillon, O. et al. Tempo and mode of genome evolution in a 50,000-generation experiment. Nature 536, 165170 (2016).

Article CAS PubMed PubMed Central Google Scholar

Lieberman, T. D. et al. Genetic variation of a bacterial pathogen within individuals with cystic fibrosis provides a record of selective pressures. Nat. Genet. 46, 8287 (2014).

Article CAS PubMed Google Scholar

Good, B. H., McDonald, M. J., Barrick, J. E., Lenski, R. E. & Desai, M. M. The dynamics of molecular evolution over 60,000 generations. Nature 551, 4550 (2017).

Article PubMed PubMed Central Google Scholar

Kumar, R. et al. Evolutionary features in the structure and function of bacterial toxins. Toxins 11, 15 (2019).

Article CAS PubMed PubMed Central Google Scholar

Whitlock, M. C. Fixation probability and time in subdivided populations. Genetics 164, 767779 (2003).

Article PubMed PubMed Central Google Scholar

Ford, C. B. et al. Use of whole genome sequencing to estimate the mutation rate of Mycobacterium tuberculosis during latent infection. Nat. Genet. 43, 482486 (2011).

Article CAS PubMed PubMed Central Google Scholar

Eldholm, V. et al. Evolution of extensively drug-resistant Mycobacterium tuberculosisfrom a susceptible ancestor in a single patient. Genome Biol. 15, 490 (2014).

Article PubMed PubMed Central Google Scholar

Walker, T. M. et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect. Dis. 13, 137146 (2013).

Article CAS PubMed PubMed Central Google Scholar

Eyre, D. W. et al. Diverse sources of C. difficile infection identified on whole-genome sequencing. N. Engl. J. Med. 369, 11951205 (2013).

Article CAS PubMed Google Scholar

He, M. et al. Emergence and global spread of epidemic healthcare-associated Clostridium difficile. Nat. Genet. 45, 109113 (2013).

Article CAS PubMed Google Scholar

Young, B. C. et al. Evolutionary dynamics of Staphylococcus aureus during progression from carriage to disease. Proc. Natl Acad. Sci. USA 109, 45504555 (2012).

Article CAS PubMed PubMed Central Google Scholar

Didelot, X., Walker, A. S., Peto, T. E., Crook, D. W. & Wilson, D. J. Within-host evolution of bacterial pathogens. Nat. Rev. Microbiol. 14, 150162 (2016).

Article CAS PubMed PubMed Central Google Scholar

Wang, J. et al. Bacillus amyloliquefaciens FH-1 significantly affects cucumber seedlings and the rhizosphere bacterial community but not soil. Sci. Rep. 11, 12055 (2021).

Article CAS PubMed PubMed Central Google Scholar

Hofgaard, I. S. et al. Inoculum potential of Fusarium spp. relates to tillage and straw management in Norwegian fields of spring oats. Front. Microbiol. 7, 556 (2016).

Article PubMed PubMed Central Google Scholar

Garud, N. R., Good, B. H., Hallatschek, O. & Pollard, K. S. Evolutionary dynamics of bacteria in the gut microbiome within and across hosts. PLOS Biol. 17, e3000102 (2019).

Article PubMed PubMed Central Google Scholar

Caugant, D. A. & Brynildsrud, O. B. Neisseria meningitidis: using genomics to understand diversity, evolution and pathogenesis. Nat. Rev. Microbiol. 18, 8496 (2020).

Article CAS PubMed Google Scholar

Montecucco, C. & Rappuoli, R. Living dangerously: how Helicobacter pylori survives in the human stomach. Nat. Rev. Mol. Cell Biol. 2, 457466 (2001).

Article CAS PubMed Google Scholar

Ely, B. Recombination and gene loss occur simultaneously during bacterial horizontal gene transfer. PLoS ONE 15, e0227987 (2020).

Article CAS PubMed PubMed Central Google Scholar

Newton, A. & Allebach, E. Gene transfer in Caulobacter crescentus: polarized inheritance of genetic markers. Genetics 80, 111 (1975).

Article CAS PubMed PubMed Central Google Scholar

Kilkenny, C., Browne, W. J., Cuthill, I. C., Emerson, M. & Altman, D. G. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 8, e1000412 (2010).

Article PubMed PubMed Central Google Scholar

Olm, M. R. et al. inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains. Nat. Biotechnol. 39, 727736 (2021).

Article CAS PubMed PubMed Central Google Scholar

Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinform. 11, 119 (2010).

Article Google Scholar

Desjardins, P. & Conklin, D. NanoDrop microvolume quantitation of nucleic acids. J. Vis. Exp. 25, 2565 (2010).

Google Scholar

Lozano-Peral, D. et al. DNA degradation in human teeth exposed to thermal stress. Sci. Rep. 11, 12118 (2021).

Article CAS PubMed PubMed Central Google Scholar

Beghini, F. et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife 10, e6508 (2021).

Read the rest here:
Native microbiome dominates over host factors in shaping the ... - Nature.com

Butler, M. S. et al. Analysis of the clinical pipeline of treatments for drug-resistant bacterial infections: despite progress, more action is needed. Antimicrob. Agents Chemother. 66, e0199121 (2022).

Article PubMed Google Scholar

Martnez-Garca, L., Gonzlez-Alba, J. M., Baquero, F., Cantn, R. & Galn, J. C. Ceftazidime is the key diversification and selection driver of VIM-type carbapenemases. mBio 9, e02109e02117 (2018).

Article PubMed PubMed Central Google Scholar

Baquero, F. et al. Allogenous selection of mutational collateral resistance: old drugs select for new resistance within antibiotic families. Front. Microbiol. 12, 757833 (2021).

Article PubMed PubMed Central Google Scholar

Rajer, F., Allander, L., Karlsson, P. A. & Sandegren, L. Evolutionary trajectories toward high-level -lactam/-lactamase inhibitor resistance in the presence of multiple -lactamases. Antimicrob. Agents Chemother. 66, e0029022 (2022).

Article PubMed Google Scholar

Poirel, L., Sadek, M., Kusaksizoglu, A. & Nordmann, P. Co-resistance to ceftazidimeavibactam and cefiderocol in clinical isolates producing KPC variants. Eur. J. Clin. Microbiol. Infect. Dis. 41, 677680 (2022).

Article CAS PubMed Google Scholar

Silver, L. L. Challenges of antibacterial discovery. Clin. Microbiol. Rev. 24, 71109 (2011).

Article CAS PubMed PubMed Central Google Scholar

Baker, S. J., Payne, D. J., Rappuoli, R. & De Gregorio, E. Technologies to address antimicrobial resistance. Proc. Natl Acad. Sci. USA 115, 1288712895 (2018).

Article CAS PubMed PubMed Central Google Scholar

Payne, D. J., Gwynn, M. N., Holmes, D. J. & Pompliano, D. L. Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nat. Rev. Drug Discov. 6, 2940 (2007).

Article CAS PubMed Google Scholar

Tommasi, R., Brown, D. G., Walkup, G. K., Manchester, J. I. & Miller, A. A. ESKAPEing the labyrinth of antibacterial discovery. Nat. Rev. Drug Discov. 14, 529542 (2015).

Article CAS PubMed Google Scholar

Brotz-Oesterhelt, H. & Sass, P. Postgenomic strategies in antibacterial drug discovery. Future Microbiol. 5, 15531579 (2010).

Article PubMed Google Scholar

Lewis, K. Platforms for antibiotic discovery. Nat. Rev. Drug Discov. 12, 371387 (2013).

Article CAS PubMed Google Scholar

Muoz, K. A. & Hergenrother, P. J. Facilitating compound entry as a means to discover antibiotics for Gram-negative bacteria. Acc. Chem. Res. 54, 13221333 (2021).

Article PubMed PubMed Central Google Scholar

Zhao, S. et al. Defining new chemical space for drug penetration into Gram-negative bacteria. Nat. Chem. Biol. 16, 12931302 (2020).

Article CAS PubMed PubMed Central Google Scholar

Sadybekov, A. V. & Katritch, V. Computational approaches streamlining drug discovery. Nature 616, 673685 (2023).

Article CAS PubMed Google Scholar

Lluka, T. & Stokes, J. M. Antibiotic discovery in the artificial intelligence era. Ann. NY Acad. Sci. 1519, 7493 (2023).

Article CAS PubMed Google Scholar

Lyu, J. et al. Ultra-large library docking for discovering new chemotypes. Nature 566, 224229 (2019).

Article CAS PubMed PubMed Central Google Scholar

Mobegi, F. M. et al. From microbial gene essentiality to novel antimicrobial drug targets. BMC Genomics 15, 958 (2014).

Article PubMed PubMed Central Google Scholar

Agarwal, P. & Searls, D. B. Can literature analysis identify innovation drivers in drug discovery? Nat. Rev. Drug Discov. 8, 865 (2009).

Article CAS PubMed PubMed Central Google Scholar

Lange, R. P., Locher, H. H., Wyss, P. C. & Then, R. L. The targets of currently used antibacterial agents: lessons for drug discovery. Curr. Pharm. Des. 13, 31403154 (2007).

Article CAS PubMed Google Scholar

Theuretzbacher, U. & Piddock, L. J. V. Non-traditional antibacterial therapeutic options and challenges. Cell Host Microbe 26, 6172 (2019).

Article CAS PubMed Google Scholar

Frearson, J. A., Wyatt, P. G., Gilbert, I. H. & Fairlamb, A. H. Target assessment for antiparasitic drug discovery. Trends Parasitol. 23, 589595 (2007).

Article CAS PubMed PubMed Central Google Scholar

Streker, K. et al. In vitro and in vivo validation of ligA and tarI as essential targets in Staphylococcus aureus. Antimicrob. Agents Chemother. 52, 44704474 (2008).

Article CAS PubMed PubMed Central Google Scholar

Rancati, G., Moffat, J., Typas, A. & Pavelka, N. Emerging and evolving concepts in gene essentiality. Nat. Rev. Genet. 19, 3449 (2018).

Article CAS PubMed Google Scholar

Bosch-Guiteras, N. & van Leeuwen, J. Exploring conditional gene essentiality through systems genetics approaches in yeast. Curr. Opin. Genet. Dev. 76, 101963 (2022).

Article CAS PubMed Google Scholar

Zhang, Z. & Ren, Q. Why are essential genes essential? the essentiality of Saccharomyces genes. Microb. Cell 2, 280287 (2015).

Article CAS PubMed PubMed Central Google Scholar

Chessher, A. Evaluating the suitability of essential genes as targets for antibiotic screening assays using proteomics. Protein Cell 3, 57 (2012).

Article PubMed PubMed Central Google Scholar

Editorial. Putting gene essentiality into context. Nat. Rev. Genet. 19, 1 (2017).

Google Scholar

Chaudhary, A. S., Chen, W., Jin, J., Tai, P. C. & Wang, B. SecA: a potential antimicrobial target. Future Med. Chem. 7, 9891007 (2015).

Article CAS PubMed Google Scholar

Emmerich, C. H. et al. Improving target assessment in biomedical research: the GOT-IT recommendations. Nat. Rev. Drug Discov. 20, 6481 (2021).

Article CAS PubMed Google Scholar

Murima, P., McKinney, J. D. & Pethe, K. Targeting bacterial central metabolism for drug development. Chem. Biol. 21, 14231432 (2014).

Article CAS PubMed Google Scholar

Hedstrom, L. The bare essentials of antibiotic target validation. ACS Infect. Dis. 3, 24 (2017).

Article CAS PubMed Google Scholar

Hogan, A. M. & Cardona, S. T. Gradients in gene essentiality reshape antibacterial research. FEMS Microbiol. Rev. 46, fuac005 (2022).

Article PubMed PubMed Central Google Scholar

Bergmiller, T., Ackermann, M. & Silander, O. K. Patterns of evolutionary conservation of essential genes correlate with their compensability. PLoS Genet. 8, e1002803 (2012).

Article CAS PubMed PubMed Central Google Scholar

Dunman, P. M. & Tomaras, A. P. Translational deficiencies in antibacterial discovery and new screening paradigms. Curr. Opin. Microbiol. 27, 108113 (2015).

Article CAS PubMed PubMed Central Google Scholar

Rosconi, F. et al. A bacterial pan-genome makes gene essentiality strain-dependent and evolvable. Nat. Microbiol. 7, 15801592 (2022).

Article CAS PubMed PubMed Central Google Scholar

Kaur, H., Kalia, M. & Taneja, N. Identification of novel non-homologous drug targets against Acinetobacter baumannii using subtractive genomics and comparative metabolic pathway analysis. Microb. Pathog. 152, 104608 (2021).

Article CAS PubMed Google Scholar

Uddin, R., Masood, F., Azam, S. S. & Wadood, A. Identification of putative non-host essential genes and novel drug targets against Acinetobacter baumannii by in silico comparative genome analysis. Microb. Pathog. 128, 2835 (2019).

Article CAS PubMed Google Scholar

Ramos, P. I. P. et al. An integrative, multi-omics approach towards the prioritization of Klebsiella pneumoniae drug targets. Sci. Rep. 8, 10755 (2018).

Article PubMed PubMed Central Google Scholar

Rafiq, H. et al. A computational subtractive genome analysis for the characterization of novel drug targets in Klebsiella pneumoniae strain PittNDM01. Microb. Pathog. 146, 104245 (2020).

Article CAS PubMed Google Scholar

Uddin, R. & Jamil, F. Prioritization of potential drug targets against P. aeruginosa by core proteomic analysis using computational subtractive genomics and proteinprotein interaction network. Comput. Biol. Chem. 74, 115122 (2018).

Article CAS PubMed Google Scholar

Nazarshodeh, E., Marashi, S. A. & Gharaghani, S. Structural systems pharmacology: a framework for integrating metabolic network and structure-based virtual screening for drug discovery against bacteria. PLoS ONE 16, e0261267 (2021).

Article CAS PubMed PubMed Central Google Scholar

Bakheet, T. M. & Doig, A. J. Properties and identification of antibiotic drug targets. BMC Bioinformatics 11, 195 (2010).

Article PubMed PubMed Central Google Scholar

Bosch, B. et al. Genome-wide gene expression tuning reveals diverse vulnerabilities of M. tuberculosis. Cell 184, 45794592.e24 (2021).

Article CAS PubMed PubMed Central Google Scholar

Hawkins, J. S. et al. Mismatch-CRISPRi reveals the co-varying expressionfitness relationships of essential genes in Escherichia coli and Bacillus subtilis. Cell Syst. 11, 523535.e9 (2020).

Article CAS PubMed PubMed Central Google Scholar

Luo, H., Gao, F. & Lin, Y. Evolutionary conservation analysis between the essential and nonessential genes in bacterial genomes. Sci. Rep. 5, 13210 (2015).

Article CAS PubMed PubMed Central Google Scholar

Arun, P. V. P. S. et al. Identification and functional analysis of essential, conserved, housekeeping and duplicated genes. FEBS Lett. 590, 14281437 (2016).

Article CAS PubMed Google Scholar

Du, W. et al. Two active forms of UDP-N-acetylglucosamine enolpyruvyl transferase in Gram-positive bacteria. J. Bacteriol. 182, 41464152 (2000).

Article CAS PubMed PubMed Central Google Scholar

Naz, S., Ngo, T., Farooq, U. & Abagyan, R. Analysis of drug binding pockets and repurposing opportunities for twelve essential enzymes of ESKAPE pathogens. PeerJ 5, e3765e3765 (2017).

Article PubMed PubMed Central Google Scholar

Klahn, P. & Bronstrup, M. New structural templates for clinically validated and novel targets in antimicrobial drug research and development. Curr. Top. Microbiol. Immunol. 398, 365417 (2016).

CAS PubMed Google Scholar

Miranda, R. R., Parthasarathy, A. & Hudson, A. O. Exploration of chemical biology approaches to facilitate the discovery and development of novel antibiotics. Front. Trop. Dis. 3, 845469 (2022).

Article Google Scholar

See the original post:
Unrealized targets in the discovery of antibiotics for Gram-negative ... - Nature.com

Credits

Mark C. Taylor is a professor of religion at Columbia University.

This essay is adapted from his forthcoming book: After the Human: A Philosophy for the Future.

Do you think human beings are the last stage in evolution? If not, what comes next?

I do not think human beings are the last stage in the evolutionary process. Whatever comes next will be neither simply organic nor simply machinic but will be the result of the increasingly symbiotic relationship between human beings and technology.

Bound together as parasite/host, neither people nor technologies can exist apart from the other because they are constitutive prostheses of each other. Such an interrelation is not unique to human beings. As the physiologist J. Scott Turner writes in The Extended Organism: Animal-built structures are properly considered organs of physiology, in principle no different from, and just as much a part of the organism as kidneys, heart, lungs or livers. This is true for termites, for example, who form a single organism in symbiosis with their nests. The extended body of the organism is created by the extended mind of the colony.

If we have an expanded understanding of body and mind, and if nature and technology are inseparably entangled, then the notion of artificiality is misleading. So-called artificial intelligence is the latest extension of the emergent process through which life takes ever more diverse and complex forms.

Our consideration of quantum phenomena, mindful bodies, relational ecology, and plant and animal cognition has revealed that we are surrounded by and entangled with all kinds of alternative intelligences. AI is another form of alternative intelligence. Critics will argue that what makes AI different is that it has been deliberately created by human beings. However, all organisms both shape and are shaped by their expanding bodies and minds. Instead of being obsessed with the prospect of creating machines whose operation is indistinguishable from human cognition, it is more important to consider how AI is different from human intelligence. The question should not be: Can AI do what humans can do? But rather: What can AI do that humans cannot do?

What is needed is a non-anthropocentric form of artificial intelligence. If humanity is to live on, AI must become smarter than the people who have created it. Why should we be preoccupied with aligning superintelligence with human values when human values are destroying the Earth, without which humans and many other forms of life cannot survive?

With the growing entanglement of the biosphere and the technosphere, further symbiogenesis is the only way to address the very real existential threat we face. But it is all too easy to wax optimistic about the salvific benefits of technology without being specific. Here I want to suggest four trajectories that will be increasingly important for the symbiotic relationship between humans and machines: neuroprosthetics, biobots, synthetic biology and organic-relational AI.

Whatever comes after the human will be neither simply organic nor machinic but the result of the increasingly symbiotic relationship between human beings and technology.

We live during a time when dystopian dread has been weaponized to create paralyzing despair that leaves many people especially the young hopeless. Without underestimating the actual and possible detrimental effects of rapid technological change, it is important not to let these dark visions overshadow the remarkable benefits many of these technologies bring.

As a long-time Type 1 diabetic, my life depends on a digital prosthesis I wear on my belt 24/7/365, which operates by artificial intelligence and is connected to the internet. Just as the Internet of Bodies creates unprecedented possibilities for monitoring and treating bodily ailments, so the Internet of Things connects smart devices wired to global networks that augment intelligence by expanding the mind. While critics and regulators of recent innovations attempt to distinguish the technologies used for therapy, which are acceptable, from technologies used for enhancement, which are unacceptable, the line between these alternative applications is fuzzy at best. What starts as treatment inevitably becomes enhancement.

Neither neuroprosthetics nor cognitive augmentation is new. Writing, after all, is a mnemonic technology that enhances the mind. In modern times, we have been enabled to archive and access memories through personal devices. Most recently, technological innovations have taken cognitive enhancement to another level: brain implants, for example, have been around since at least 2006, and entrepreneurs like Elon Musk (who founded Neuralink to create symbiosis with artificial intelligence) aim to establish embodied human-machine interfaces. Increasing possibilities for symbiotic relations between computers and brains will lead to alternative forms of intelligence that are neither human nor machinic, but something in between.

So-called artificial intelligence is the latest extension of the emergent process through which life takes ever more diverse and complex forms.

In recent years, there has been a revolution in robotics as the result of developments in nanotechnology and the refinement of large language models like ChatGPT. Individual as well as swarms of nanobots might one day be implanted in the body and used for diagnostic and therapeutic purposes, potentially delivering drugs and repairing tissue. Rather than working through the entire body, nanobots might target the precise location where a drug is needed and regulate its delivery.

The most noteworthy deployment in nanotechnology to date is its use in vaccines, including the Covid vaccines. As a group of microbiology and pharmacology experts wrote in a 2021 paper, Nanotechnology has played a significant role in the success of these vaccines; the emergency use authorization that allowed the rapid development and testing of this technology is a major milestone and showcases the immense potential of nanotechnology for vaccine delivery and for fighting against future pandemics. Nanotechnology research and development are in the very early stages but are developing rapidly. As they progress, not only will bodies become more mindful, but it will be increasingly difficult to distinguish the natural from the artificial.

While nanobots are implanted in the body and operate at the molecular level, other robots are becoming both increasingly autonomous and able to think and act in ways that are more human-like. Kevin Roose reported in the New York Times that Googles latest robot RT-2 can interpret images and analyze the surrounding world. It does so by translating the robots movements into a series of numbers a process called tokenizing and incorporating those tokens into the same training data as the language model. Eventually, just as ChatGPT or Bard learns to guess what words should come next in a poem or a history essay, RT-2 can learn to guess how a robots arm should move to pick up a ball or throw an empty soda can into the recycling bin. Thus, rather than programming a robot to perform a specific task, it is possible to give the robot instructions for the task to be performed and to let the machine figure out how to do it.

Building on these recent advances, Hod Lipson, the director of the Creative Machines Lab at Columbia University, is taking robotic research to the next level, building robots thatcreateandare creative. His research is inspired from biology, and he is searching for new biological concepts for engineering and new engineering insights into biology.

It will be increasingly difficult to distinguish the natural from the artificial.

Lipsons ultimate goal is to create robots that not only can reason but also are conscious and self-aware. Defining consciousness as the ability to imagine yourself in the future, he confidently predicts that eventually these machines will be able to understand what they are, and what they think. As cognitive skills enabled by generative AI become more sophisticated, physical movements and activities will become more natural. With these new skills, robots might have the agility to navigate in their surroundings as effectively as humans.

Science and art meet in biobots. David Hanson is the founder and CEO of Hanson Robotics, a Hong Kong-based company founded in 2013, a musician who has collaborated with David Byrne of the Talking Heads, and a sculptor. His best-known work is a humanoid smart robot named Sophia who, he says, personifies our dreams for the future of AI. As a unique combination of science, engineering and artistry, Sophia is simultaneously a human-crafted science fiction character depicting the future of AI and robotics, and a platform for advanced robotics and AI research. She is the first robot citizen and the first robot Innovation Ambassador for the United Development Program.

Speaking for herself, Sophia adds, In some ways, I am a human-crafted science-fiction character depicting where AI and robotics are heading. In other ways, I am real science, springing from the serious engineering and science research and accomplishments of an inspired team of roboticists and AI scientists and designers.

Sophia is so realistic that people have fallen in love and proposed marriage to her. The writer Sue Halpern reports that In 2017, the government of Saudi Arabia gave Sophia citizenship, making it the first state to grant personhood to a machine. The response to Sophia suggests that as robots become more proficient and are integrated into everyday life, they will become less uncanny. The theory of the uncanny valley, perhaps, might turn out to be wrong.

Nowhere are the biosphere and the technosphere more closely interrelated than in synthetic biology. This field includes disciplines ranging from various branches of biology, chemistry, physics, neurology and computer engineering. Michael Levin and his colleagues at the Allen Discovery Center of Tufts University biologists, computer scientists and engineers have created xenobots, which are biological robots that were produced from embryonic skin and muscle cells from an African clawed frog (Xenopus laevis). These cells are manually manipulated in a sculpting process guided by algorithms. Like Sophia, xenobots are sculptures that complicate the boundary between organism and machine. As Levin and his colleagues wrote in 2020:

Living systems are more robust, diverse, complex and supportive of human life than any technology yet created. However, our ability to create novel lifeforms is currently limited to varying existing organisms or bioengineering organoids in vitro. Here we show a scalable pipeline for creating functional novel lifeforms: AI methods automatically design diverse candidate lifeforms in silico to perform some desired function, and transferable designs are then created using a cell-based construction toolkit to realize living systems with predicted behavior. Although some steps in this pipeline still require manual intervention, complete automation in the future would pave the way for designing and deploying living systems for a wide range of functions.

Xenobots use evolutionary algorithms to modify the computational capacity of cells to create the possibility of novel functions and even new morphologies. Aggregates of cells display novel functions that bear little resemblance to existing organs or organisms. Through a process of trial and error, evolutionary algorithms design cells harvested from skin and heart muscle cells to perform specific tasks like walking, swimming and pushing other entitles. Collections of xenobots display swarming behaviors characteristic of other emergent complex adaptive systems; they can self-assemble, self-organize, self-replicate and self-repair. Levin envisions multiple applications of this biomechanic technology from using self-renewing biocompatible biological robots to cure living systems to creating materials with less harmful effects, delivering drugs internally that repair organs and even growing organs that can be transplanted in humans.

Machines are becoming more like people and people are becoming more like machines.

In 2021, Levin and his colleagues published a follow-up study, in which he reported on a successful experiment in which he created xenobots that independently developed their shape and began to function on their own:

These xenobots exhibit coordinated locomotion via cilia present on their surface. These cilia arise through normal tissue patterning and do not require complicated construction methods or genomic editing, making production amenable to high-throughput projects. The biological robots arise by cellular self-organization and do not require scaffolds or microprinting; the amphibian cells are highly amenable to surgical, genetic, chemical and optical stimulation during the self-assembly process. We show that the xenobots can navigate aqueous environments in diverse ways, heal after damage and show emergent group behaviors.

This generation of xenobots exhibits bottom-up swarming behavior, which, like all emergent complex adaptive networks, is the result of the interaction of multiple individual components that are closely interrelated.

Algorithms program sensation and memory into the xenobots, which communicate with each other through biochemical and electrical signaling. The skin cells use the same electrical processes used in the brains neural network. As Philip Ball writes in Quanta Magazine, Intercellular communications create a sort of code that imprints a form, and cells can sometimes decide how to arrange themselves more or less independently of their genes. In other words, the genes provide the hardware, in the form of enzymes and regulatory circuits for controlling their production. But the genetic input doesnt in itself specify the collective behavior of cell communities.

It is important to stress that these xenobots are autonomous. As Levin and his colleagues conclude their 2021 paper: The computational modeling of unexpected, emergent properties at multiple scales and the apparent plasticity of cells with wild-type genomes to cooperate toward the construction of various functional body architectures offer a very potent synergy. Like superorganisms and superintelligence, the behavior of entangled xenobots is, in an important sense, out of control. While this indeterminacy creates uncertainty, it is also the source of evolutionary novelty. Eva Jablonka, who is an evolutionary biologist at Tel Aviv University, believes that xenobots are a new type of organism, one defined by what it does rather than to what it belongs developmentally or evolutionarily.

While Levin uses computational technology to create and modify biological organisms, the German neurobiologist Peter Robin Hiesinger uses biological organisms to model computational processes by creating algorithms that evolve. This work involves nothing less than developing a new form of artificial intelligence.

According to the pioneering work by James Watson, Francis Crick and other early DNA researchers, a genome functions as a program that serves as the blueprint for the production of an organism. Summarizing this process, Hiesinger raises questions about the accuracy of the metaphor code. Genes encode proteins, proteins encode an interaction network, etc. But what does encode mean yet again? he writes in his 2021 book The Self-Assembling Brain. He continues:

The gene contains information for the primary amino acids sequence, but we cannot read the protein structure in the DNA. The proteins arguably contain information about their inherent ability to physically interact with other proteins, but not when and what interactions actually happen. The next level up, what are neuronal properties? A property like neuronal excitability is shaped by the underlying protein interaction network, e.g., ion channels that need to be anchored at the right place in the membrane. But neuronal excitability is also shaped by the physical properties of the axon, the ion distribution and many other factors, all themselves a result of the actions of proteins and their networks.

It becomes clear that a one-way model for gene-protein interaction is vastly oversimplified. The genotype does not only determine the phenotype, but the phenotype and its relation to the environment also alters the genotype. Hiesinger explains that this reciprocal relationship is even more complicated. Rather than a prescribed program, the genome is a complicated relational network in which both genes and proteins contain the information required to generate the organism. The information of the genes is in part the result of the interactions that occur in a network of proteins.

The reciprocal gene-protein interaction changes the understanding of the genome. The genome is not a prescribed program that determines the structure and operation of the organism. The genome is not fixed in advance but evolves in relation to the information created by the interactions of the proteins it partially produces, which, in turn, reconfigure the genome.

The brain and its development, for example, are not completely programmed in advance but coevolve through a complicated network of connections. Hiesinger uses the illuminating example of navigating city streets to explain the process of the brains self-assembling of neuronal circuits:

How are such connections made during the brains development? You can imagine yourself trying to make a connection by navigating the intricate network of city streets. Except, you wont get far, at least not if you are trying to understand brain development. There is a problem with that picture, and it is this: Where do the streets come from? Most connections in the brain are not made by navigating existing streets, but by navigating streets under construction. For the picture to make sense, you would have to navigate at the time the city is still growing, adding street by street, removing and modifying old ones in the process, all while traffic is part of city life. The map changes as you are changing your position in it, and you will only ever arrive if the map changes in interaction with your own movements in it. The development of brain wiring is a story of self-assembly, not a global positioning system.

The successful creation of evolving networks and algorithms would create an even closer symbiotic relationship between the biosphere and the technosphere.

In this model, there is no blueprint for brain connectivity encoded in the genes:

Genetic information allows brains to grow. Development progresses in time and requires energy. Step by step, the developing brain finds itself in changing configurations. Each configuration serves as a new basis for the next step in the growth process. At each step, bits of the genome are activated to produce gene products that themselves change what parts of the genome will be activated next a continuous feedback process between genome and its products. Rather than dealing with endpoint information, the information to build the brain unfolds with time. Remarkably, there may be no other way to read the genetic information than to run the program.

Hiesinger argues that this understanding of the brains self-assembling neural networks points to an alternative model of not-so-artificial intelligence that differs from both symbolic AI and artificial neural networks (ANNs), as well as their extension in generative AI. The genome functions as an algorithm or as a network of entangled algorithms, which does not preexist the organ or organism but coevolves along with it what it both produces and, in turn, is produced by it.

In other words, neither the genome (algorithm) nor the connectivity of the network is fixed in advance of their developmental process. The brain doesnt come into being fully wired with an empty network, all ready to run, just without information, Hiesinger writes. As the brain grows, the wiring precision develops. This creates a feedback loop that never stops and, therefore, the algorithmic growth of biological networks is continuous.

In symbolic AI, a fixed network architecture facilitates the application of fixed rules (algorithms) in a top-down fixed sequence to externally provided data. Artificial neural networks, by contrast, do not start with prescribed algorithms but generate patterns and rules in a bottom-up process that allows for algorithmic change. Relative weights change, but the network architecture does not.

Hiesinger proposes that the self-assembly of the brains neural network provides a more promising model for AI than either symbolic AI or ANNs. The successful creation of evolving networks and algorithms would create an even closer symbiotic relationship between the biosphere and the technosphere.

One of the concerns about developing organic AI is its unpredictability and the uncertainty it creates. Human control of natural, social and cultural processes is, however, an illusion created by the seemingly insatiable will to mastery that has turned destructive. As Hiesinger correctly claims, An artificial intelligence need not be humanlike, to be as smart (or smarter than) a human. Non-anthropocentric AI would not be merely an imitation of human intelligence, but would be as different from our thinking as fungi, dog and crow cognition is from human cognition.

Machines are becoming more like people and people are becoming more like machines. Organism and machine? Organism or machine? Neither organism nor machine? Evolution is not over; something new, something different, perhaps infinitely and qualitatively different, is emerging. Who would want the future to be the endless repetition of the past?

Read the original:
How Biotech And AI Are Transforming The Human - Noema Magazine

Pictured: Alexandria Forbes/Getty Images, modified by Nicole Bean for BioSpace

One of the most memorable events of Alexandria Forbes life is being four years old and incubating a clutch of lizard eggs in her kitchen. A naturally inquisitive child, Forbes found the eggs outside and kept them in a paper cup until they hatched.

Ive always had this very strong curiosity and acute observation of life, and that probably underlies my interest in biology as opposed to physics or engineering, Forbes said.

The MeiraGTx CEO was born in the West Indies and quickly developed a love for exploring the regions beaches and gardens. What began as a penchant for observation would eventually lead Forbes to found the clinical-stage gene therapy company in 2015.

This year, Forbes was named to Forbes Medias 50 Over 50 class in the innovation category, and she said she was delighted to be recognized both as a woman scientist and as someone who didnt have a standard career path.

I had a very successful academic career, but then I did something else, and then I had a very successful financial career, but then I did something else, Forbes said.

Leading a biotechnology company was not something Forbes anticipated.

While attending school in the United Kingdom, Forbes excelled in every subject, including science. Despite this, she didnt plan on pursuing the sciences until her teachers encouraged her to take science A-levels. Forbes took A-levels in biology and chemistry instead of English and history as shed originally planned, and later matriculated to the natural sciences program at the University of Cambridge. At that time, the human genome had not been sequenced yet.

I decided in my last year [of undergrad] that the thing I wanted to do was understand the way cells work, Forbes said.

Deciding what to do after graduation was a different story. Englands financial industry was booming in the late 1980s, and Forbes considered becoming a trader before ultimately deciding to pursue a PhD in molecular genetics at the University of Oxford, where she researched signaling pathways in fruit flies. Her doctoral training would later come into play when she began working with MeiraGTx, particularly when it came to developing the technology the company would be built on.

After stints at Duke University, the Carnegie Institute at Johns Hopkins University and the Skirball Institute of Biomolecular Medicine at NYU Langone Medical Center, Forbes shifted from research to finance in 2000 when a former Cambridge colleague encouraged her to take an interview as a buy-side analyst. She worked as a healthcare investor at Sivik Global Healthcare from 2000 to 2008 and Meadowvale Asset Management from 2008 to 2013. She eventually became senior vice president of commercial operations at Kadmon Holdings, Inc., which provided some of the necessary resources to launch MeiraGTx.

While working in investment, Forbes managed a biotech fund, and she said she learned more about drug development from this experience than she would have learned working at a pharmaceutical company that focused on a handful of drugs. It also gave her a real understanding of risk and reward for biotech companies, she said.

You never forget why you were wrong. And that decade, I was wrong many, many times, Forbes said.

Through this work, Forbes saw an opportunity to finance a company in the growing gene therapy arena, and she founded MeiraGTx in 2015 with a vision of building a unique, end-to-end, vertically integrated approach to gene therapy, she wrote in an email to BioSpace. The company went public in 2018, a choice guided by Forbes and co-founder Richard Girouxs familiarity with public investors. MeiraGTxs initial clinical programs were selected because they already had human proof-of-concept; the company develops treatments for diseases of the salivary gland, eyes and central nervous system. It also decided to focus on localized gene therapy delivery to minimize safety and manufacturing issues.

MeiraGTx therapies rely on adeno-associated viruses, which can carry genetic material, and riboswitches, which regulate gene expression. The latter technology was invented by the company.

One of the things I learned during my PhD is that when you want a really sharp signal or a really sharp on/off, you tend not to get that by trying to switch something on, Forbes said of riboswitches. What we did is repress the repression of something, and in life, particularly in developmental biology, thats how it works.

MeiraGTx also built its own manufacturing facility, and then a second in 2021 that also functions as a quality control facility.

This planning made the companys later ventures possible. Having manufacturing in-house allowed MeiraGTx to partner with Janssen for clinical development, Forbes said, which has led to Phase III studies and the opportunity to file Biologic License Applications for the companys products. Working with other companies has also allowed MeiraGTx to refine how it approaches Investigational New Drug applications and other aspects of the regulatory process, Forbes said.

I learn every day doing this job, right? Forbes said. How to deal with people, how to deal with failure, how to deal with success, all of those thingsevery day I learn to do something new, and maybe thats why Ive done three different careers. Because I like learning new things.

Nadia Bey is a freelance writer from North Carolina. She can be reached at beynadiaa@gmail.com.

More here:
The Many Lives of Alexandria Forbes - BioSpace

Ethics approval statement

All the human non-small cell lung cancer samples were obtained from Zhejiang Cancer Hospital with informed consent. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The Ethics Committee in Clinical Research (ECCR) of The Cancer Hospital of the University of Chinese Academy of Sciences (Zhejiang Cancer Hospital) approved this retrospective study (IRB-2022-268). The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

All patients signed an informed consent approved by the institutional Review Board.

Quantitative real-timepolymerase chain reaction (qRT-PCR) was performed using 2Synergy Brands Inc (SYBR) Green (Servicebio) on the Roche LightCycler480 II RT-PCR Detection System. The relative CEP20 expression was quantified by RT-PCR, and Actin was used as an internal reference. All the reactions were triplicated and were calculated using the comparative threshold method (({2}^{{ - Delta Delta {text{C}}_{{text{t}}} }})).

Cell lysates or microtubule pellets were subjected to western blotting analysis with anti-CEP20, Actin, or GAPDH antibodies (Sigma, St Louis, MO, USA). The blots were probed with either Alexa Fluor 680 or IRDye 800-conjugated secondary antibodies and detected using the Odyssey system (LI-COR Biosciences, Lincoln, NE, USA). The uncropped immunoblot images are presented in Fig. S8.

A549 and H1299 cells were cultured in a complete DMEM (10% FBS included) medium with 5% CO2 at 37C. H226 and H520 were cultured in a complete RPMI 1640 (10% FBS included) medium with 5% CO2 at 37C. Cells were split at approximately 80% confluence by first aspiring the medium, followed by washing with preheated sterile 1PBS buffer thrice. Trypsin was given for 1min to induce cell detachment at 37C, then terminated by adding an appropriate volume of the medium. The cell mixture was transferred into a 15mL Falcon tube and dissociated to form a single-cell suspension by pipetting up and down. An appropriate volume of suspension was added to a new plate for continuous culture.

The cells were cultivated to the logarithmic growth phase and passaged the day before transfection. When cells reached 2030% confluence, Lipofectamine RNAiMAX and the siRNAs (CEP20 RNAi-1 5-ACCACTAATGTTTGTAGAATT-3 CEP20 RNAi-2 5-ATGGATGACCACCTAAGAATT-3) were diluted with DMEM (FBS free) according to the corresponding transfection system. Each dilution was incubated for 5min, mixed well, and incubated for another 20min at room temperature. After adding the mixture into corresponding groups, cells were cultured for 6h in a 5% CO2 incubator at 37C. Subsequently, the medium was replaced with complete DMEM containing 10% FBS. After 4872h of transfection, the cells were observed under a fluorescent microscope to evaluate their condition and transfection efficiency for further analysis.

A549 and H1299 cells were transfected and subsequently cultured for 48h. Then cells were evenly passage to 96-well plates with 2103 cells per well. Cultured for 24, 48, 72 and 96h, cells were added with 20 L/well MTT solution (5mg/ml, Sigma, St. Louis, MO, USA) and incubated for 4h at 37C. Then the medium was discarded and added 150 L of dimethyl sulphoxide (DMSO) (Sigma, St. Louis, MO, USA), the cell proliferation was analyzed by measuring the absorption at 490nm. Cell growth curves were depicted by Graphpad Prism 9 software.

Cells were plated on 12-well plates (200 cells per well). The cell culture medium was changed every 2days. After 2weeks, the cells were fixed with 4% paraformaldehyde for 15min, then washed 3 times by phosphate buffered solution (PBS), and dyed with crystal violet staining solution for 30min.

Cells were transfected and subsequently cultured until they reached 100% confluence. The cells were then starved overnight using a bare medium (DMEM or RPMI 1640 with no glucose or FBS). Mechanical scratching (wound) was performed manually with a pipette tip (10l), and the medium was replaced with DMEM or RPMI 1640 containing 1% FBS. Cells were imaged every 12h. It is important to note that the same area of the wound was imaged consistently across time points.

Cells were transfected following the protocol described above. After 48h of transfection, cells were starved overnight using a bare medium (DMEM or RPMI 1640 with no glucose or FBS). Subsequently, cells were trypsinized and counted, and 80,000 starved cells were resuspended in DMEM or RPMI 1640 containing 1% FBS and added to the upper chamber of Transwell inserts. The lower chamber was filled with 600l of complete DMEM or RPMI 1640 containing 10% FBS. The cells were then cultured for 4h at 37C with 5% CO2. After incubation, the transwell chambers were taken out and fixed with 4% PFA for 20min at room temperature. The inserts were stained with crystal violet for 20min, then washed for 5min each thrice. Residual non-migratory cells from the upper chamber were wiped off with a swab, while the migratory cells were counted and imaged under a microscope.

A549 cells grown on coverslips were fixed with cold methanol (20C), stained with anti-CEP20, -tubulin antibodies (Sigma, St Louis, MO, USA) for 2h at room temperature, and incubated with either Cy3-conjugated anti-mouse IgG or FITC-conjugated anti-rabbit IgG secondary antibody (Jackson ImmunoResearch) for 40min. DNA was stained with DAPI (Sigma). Finally, the mounted coverslips were analyzed by confocal fluorescence microscopy (LSM510, Zeiss).

For the cellular microtubule depolymerization assay21, A549 cells were treated with 5M nocodazole for the indicated times, and then centrifuged at 100 000g for 20min at 25C. For the cellular microtubule regrowth assay, A549 cells grown on coverslips were incubated with 5M nocodazole for 3h to depolymerize microtubules, and then carefully washed out to remove nocodazole followed by fixation at the indicated times. All cells were stained with mouse anti--tubulin primary antibody and Cy3-conjugated anti-mouse IgG secondary antibody. The coverslips were then mounted and imaged by confocal microscopy (NIKON, Tokyo, Japan). The astral length of microtubules within the region of interest were quantified using ImageJ software (Fiji, NIH). Data are expressed as means.d. and analyzed by students t-test. The supernatant and pellet fractions were collected separately and analyzed by western blotting with anti -tubulin.

Total RNA was isolated and purified using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturers instructions. The RNA amount and purity of each sample were quantified using NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA). The RNA integrity was assessed using Bioanalyzer 2100 (Agilent, CA, USA) with a RIN above 7.0 and confirmed by denaturing agarose gel electrophoresis. Poly (A) RNA was purified from 1g total RNA using Dynabeads Oligo (dT) 25-61005 (Thermo Fisher, CA, USA) with two rounds of purification. The purified poly(A) RNA was fragmented into small pieces using the Magnesium RNA Fragmentation Module (NEB, e6150, USA) at 94C for 57min. The cleaved RNA fragments were then reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen, cat. 1896649, USA). The resultant cDNA was used to synthesize U-labeled second-stranded DNAs with E. coli DNA polymerase I (NEB, m0209, USA), RNase H (NEB, m0297, USA), and dUTP solution (Thermo Fisher, R0133, USA). An A-base is then added to the blunt ends of each strand, preparing them for ligation to the indexed adapters. Each adapter contained a T-base overhang for ligating the adapter to the A-tailed fragmented DNA. Single- or dual-index adapters are ligated to the fragments, and size selection was performed with AMPureXP beads. After treatment with the heat-labile UDG enzyme (NEB, m0280, USA) to remove the U-labeled second-stranded DNAs, the ligated products are amplified using PCR. The PCR conditions were as follows: initial denaturation at 95C for 3min; 8 cycles of denaturation at 98C for 15s, annealing at 60C for 15s, and extension at 72C for 30s; and final extension at 72C for 5min. The average insert size for the final cDNA library was 30050bp. Finally, the 2150bp paired-end sequencing (PE150) was performed on an Illumina Novaseq 6000 according to the manufacturers protocol.

Raw RNA-seq data were processed using fastp (v0.20.1)25 to remove adapter sequences and reads with low sequencing quality. The remaining clean reads were aligned to the human genome (hg38) using HISAT2 software (v2.1.0)26 with default parameter settings. Transcript assembly was performed using StringTie software (v2.0)27, and expression of transcripts sharing each gene_id was quantified as Transcripts Per Million (TPM). Differential expression analysis was performed using the R package DESeq228 with a threshold of significantly differentially expressed genes set as fold change (FC)>1.5 or<0.67 and adjusted P value<0.05. Heatmaps were generated using the R package pheatmap. The Gene Ontology (GO) term and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses in current study were done by R package clusterProfiler29. Adjusted p value<0.05 was considered as statistically significant. The gene set enrichment analysis (GSEA) was performed by R package enrichplot.

The dataset GSE1980430 based on the platform of GPL570 (Affymetrix Human Genome U133 Plus 2.0 Array) containing 30 paired gene-microarray samples of human NSCLC tumor and normal tissues were downloaded from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo). The RNA-seq data of NSCLC samples were retrieved from the Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), including 513 lung adenocarcinoma (LUAD) tumor samples, 57 LUAD adjacent normal samples and 501 lung squamous cell carcinoma (LUSC) tumor samples, and 49 LUSC adjacent normal samples. The expression levels of CEP20 were extracted from these datasets, and the NSCLC samples from the TCGA database were classified into three groups based on their CEP20 expression levels: relatively high (CEP20-high, n=253), relatively low (CEP20-low, n=253), and medium (CEP20-median, n=508).

All experiment results are presented as meanstandard deviation (SD) from 3 independent experiments and showed successful reproducibility. All graphs were generated using GraphPad Prism9 (64-bit, La Jolla, CA, USA). Two-tailed unpaired t-tests (Students t-test) were used to obtain the p values. The data are presented as the meanstandard deviation. *P<0.05, **P<0.01, ***P<0.001.

See the original post here:
CEP20 promotes invasion and metastasis of non-small cell lung ... - Nature.com

Recruitment of study participants

The MCPS was established in the late 1990s following discussions between Mexican scientists at the National Autonomous University of Mexico (UNAM) and British scientists at the University of Oxford about how best to measure the changing health effects of tobacco in Mexico. These discussions evolved into a plan to establish a prospective cohort study that could investigate not only the health effects of tobacco but also those of many other factors (including factors measurable in the blood)1. Between 1998 and 2004, more than 100,000 women and 50,000 men 35years of age or older (mean age 50years) agreed to take part, were asked questions, had physical measurements taken, gave a blood sample and agreed to be tracked for cause-specific mortality. More women than men were recruited because the study visits were predominantly made during working hours when women were more likely to be at home (although visits were extended into the early evenings and at weekends to increase the proportion of men in the study).

Participants were recruited from randomly selected areas within two contiguous city districts (Coyoacn and Iztapalapa). These two districts have existed since the pre-Hispanic period and are geographically close to the ancient Aztec city of Tenochtitlan. Originally, Indigenous populations settled there, but over the centuries, the population dynamics have substantially changed. Many people from Spain, including the conqueror Hernn Corts, resided in Coyoacn while the capital of New Spain was being built over the ruins of Tenochtitlan. The modern populations of Coyoacn and Iztapalapa derive largely from the development of urban settlements and migrations from the 1950s to the 1970s. Over this period, both districts, but particularly Iztapalapa, received large numbers of Indigenous migrants from the central (Nahuas, Otomies and Purepechas), south (Mixtecos, Zapotecos and Mazatecos) and southeast (Chinantecos, Totonacas and Mayas) regions of the country.

At recruitment, a 10-ml venous EDTA blood sample was obtained from each participant and transferred to a central laboratory using a transport box chilled (410C) with ice packs. Samples were refrigerated overnight at 4C and then centrifuged (2,100g at 4C for 15min) and separated the next morning. Plasma and buffy-coat samples were stored locally at 80C, then transported on dry ice to Oxford (United Kingdom) for long-term storage over liquid nitrogen. DNA was extracted from buffy coat at the UK Biocentre using Perkin Elmer Chemagic 360 systems and suspended in TE buffer. UV-VIS spectroscopy using Trinean DropSense96 was used to determine yield and quality, and samples were normalized to provide 2g DNA at 20ngl1 concentration (2% of samples provided a minimum 1.5g DNA at 10ngl1 concentration) with a 260:280nm ratio of >1.8 and a 260:230nm ratio of 2.02.2.

Genomic DNA samples were transferred to the Regeneron Genetics Center from the UK Biocentre and stored in an automated sample biobank at 80C before sample preparation. DNA libraries were created by enzymatically shearing DNA to a mean fragment size of 200bp, and a common Y-shaped adapter was ligated to all DNA libraries. Unique, asymmetric 10bp barcodes were added to the DNA fragment during library amplification to facilitate multiplexed exome capture and sequencing. Equal amounts of sample were pooled before overnight exome capture, with a slightly modified version of IDTs xGenv1 probe library; all samples were captured on the same lot of oligonucleotides. The captured DNA was PCR amplified and quantified by quantitative PCR. The multiplexed samples were pooled and then sequenced using 75bp paired-end reads with two 10bp index reads on an Illumina NovaSeq 6000 platform on S4 flow cells. A total of 146,068 samples were made available for processing. We were unable to process 2,628 samples, most of which failed QC during processing owing to low or no DNA being present. A total of 143,440 samples were sequenced. The average 20 coverage was 96.5%, and 98.7% of the samples were above 90%.

Of the 143,440 samples sequenced, 2,394 (1.7%) did not pass one or more of our QC metrics and were subsequently excluded. Criteria for exclusion were as follows: disagreement between genetically determined and reported sex (n=1,032); high rates of heterozygosity or contamination (VBID>5%) (n=249); low sequence coverage (less than 80% of targeted bases achieving 20 coverage) (n=29); genetically identified sample duplicates (n=1,062 total samples); WES variants discordant with the genotyping chip (n=8); uncertain linkage back to a study participant (n=259); and instrument issue at DNA extraction (n=6). The remaining 141,046 samples were then used to compile a project-level VCF (PVCF) for downstream analysis using the GLnexus joint genotyping tool. This final dataset contained 9,950,580 variants.

Approximately 250ng of total DNA was enzymatically sheared to a mean fragment size of 350bp. Following ligation of a Y-shaped adapter, unique, asymmetric 10bp barcodes were added to the DNA fragments with three cycles of PCR. Libraries were quantified by quantitative PCR, pooled and then sequenced using 150bp paired-end reads with two 10bp index reads on an Illumina NovaSeq 6000 platform on S4 flow cells. A total of 10,008 samples were sequenced. This included 200 motherfatherchild trios and 3more extended pedigrees. The rest of the samples were chosen to be unrelated to third degree or closer and enriched for parents of nuclear families. The average mean coverage was 38.5 and 99% of samples had mean coverages of >30, and all samples were above 27.

Of the 10,008 samples that were whole-genome sequenced, 58 (0.6%) did not pass one or more of our QC metrics and were subsequently excluded. Reasons for exclusion were as follows: disagreement between genetically determined and reported sex (n=16); high rates of heterozygosity or contamination (VBID>5%) (n=10); genetically identified sample duplicates (n=19 total samples); and uncertain linkage back to a study participant (n=14). The remaining 9,950 samples were then used to compile a PVCF for downstream analysis using the GLnexus joint genotyping tool. This final dataset contained 158,464,363 variants.

The MCPS WES and WGS data were reference-aligned using the OQFE protocol35, which uses BWA MEM to map all reads to the GRCh38 reference in an alt-aware manner, marks read duplicates and adds additional per-read tags. The OQFE protocol retains all reads and original quality scores such that the original FASTQ is completely recoverable from the resulting CRAM file. Single-sample variants were called using DeepVariant (v.0.10.0) with default WGS parameters or custom exome parameters35, generating a gVCF for each input OQFE CRAM file. These gVCFs were aggregated and joint-genotyped using GLnexus (v.1.3.1). All constituent steps of this protocol were executed using open-source software.

Similar to other recent large-scale sequencing efforts, we implemented a supervised machine-learning algorithm to discriminate between probable low-quality and high-quality variants8,12. In brief, we defined a set of positive control and negative control variants based on the following criteria: (1) concordance in genotype calls between array and exome-sequencing data; (2) transmitted singletons; (3) an external set of likely high quality sites; and (4) an external set of likely low quality sites. To define the external high-quality set, we first generated the intersection of variants that passed QC in both TOPMed Freeze8 and gnomADv.3.1 genomes. This set was additionally restricted to 1000 genomes phase1 high-confidence SNPs from the 1000Genomes project36 and gold-standard insertions and deletions from the 1000Genomes project and a previous study37, both available through the GATK resource bundle (https://gatk.broadinstitute.org/hc/en-us/articles/360035890811-Resource-bundle). To define the external low-quality set, we intersected gnomADv3.1 fail variants with TOPMed Freeze8 Mendelian or duplicate discordant variants. Before model training, the control set of variants were binned by allele frequency and then randomly sampled such that an equal number of variants were retained in the positive and negative labels across each frequency bin. A support vector machine using a radial basis function kernel was then trained on up to 33 available site quality metrics, including, for example, the median value for allele balance in heterozygote calls and whether a variant was split from a multi-allelic site. We split the data into training (80%) and test (20%) sets. We performed a grid search with fivefold cross-validation on the training set to identify the hyperparameters that returned the highest accuracy during cross-validation, which were then applied to the test set to confirm accuracy. This approach identified a total of 616,027 WES and 22,784,296 WGS variants as low-quality (of which 161,707 and 104,452 were coding variants, respectively). We further applied a set of hard filters to exclude monomorphs, unresolved duplicates, variants with >10% missingness, 3 mendel errors (WGS only) or failed HardyWeinberg equilibrium (HWE) with excess heterozgosity (HWE P<11030 and observed heterozygote count of >1.5 expected heterozygote count), which resulted in a dataset of 9,325,897 WES and 131,851,586 WGS variants (of which 4,037,949 and 1,460,499 were coding variants, respectively).

Variants were annotated as previously described38. In brief, variants were annotated using Ensembl variant effect predictor, with the most severe consequence for each variant chosen across all protein-coding transcripts. In addition, we derived canonical transcript annotations based on a combination of MANE, APPRIS and Ensembl canonical tags. MANE annotation was given the highest priority followed by APPRIS. When neither MANE nor APPRIS annotation tags were available for a gene, the canonical transcript definition of Ensembl was used. Gene regions were defined using Ensembl release 100. Variants annotated as stop gained, start lost, splice donor, splice acceptor, stop lost or frameshift, for which the allele of interest was not the ancestral allele, were considered predicted loss-of-function variants. Five annotation resources were utilized to assign deleteriousness to missense variants: SIFT; PolyPhen2 HDIV and PolyPhen2 HVAR; LRT; and MutationTaster. Missense variants were considered likely deleterious if predicted deleterious by all five algorithms, possibly deleterious if predicted deleterious by at least one algorithm and likely benign if not predicted deleterious by any algorithm.

Samples were genotyped using an Illumina Global Screening Array (GSA) v.2 beadchip according to the manufacturers recommendations. A total of 146,068 samples were made available for processing, of which 145,266 (99.5%) were successfully processed. The average genotype call rate per sample was 98.4%, and 98.4% of samples had a call rate above 90%. Of the 145,266 samples that were genotyped, 4,435 (3.1%) did not pass one or more of our QC metrics and were subsequently excluded. Reasons for exclusion were as follows: disagreement between genetically determined and reported sex (n=1,827); low-quality samples (call rates below 90%) (n=2,276); genotyping chip variants discordant with exome data (n=44); genetically identified sample duplicates (n=1,063 total samples); uncertain linkage back to a study participant (n=268); and sample affected by an instrument issue at DNA extraction (n=6). The remaining 140,831 samples were then used to compile a PVCF for downstream analysis. This dataset contained 650,380 polymorphic variants.

The input array data from the RGC Sequencing Laboratory consisted of 140,831 samples and 650,380 variants and were passed through the following QC steps: checks for consistency of genotypes in sex chromosomes (steps14); sample-level and variant-level missingness filters (steps 5 and 6); the HWE exact test applied to a set of 81,747 third-degree unrelated samples, which were identified from the initial relatedness analysis using Plink and Primus (step7); setting genotypes with Mendelian errors in nuclear families to missing (step8); and a second round of steps57 (step9). Plink commands associated with each step are displayed in column2 (Supplementary Table 9). The final post-QC array data consisted of 138,511 samples and 559,923 variants.

We used Shapeit (v.4.1.3; https://odelaneau.github.io/shapeit4) to phase the array dataset of 138,511 samples and 539,315 autosomal variants that passed the array QC procedure. To improve the phasing quality, we leveraged the inferred family information by building a partial haplotype scaffold on unphased genotypes at 1,266 trios from 3,475 inferred nuclear families identified (randomly selecting one offspring per family when there was more than one). We then ran Shapeit one chromosome at a time, passing the scaffold information with the --scaffold option.

We separately phased the support-vector-machine-filtered WES and WGS datasets onto the array scaffold. The phased WGS data constitute the MCPS10k reference panel. For the WGS phasing, we used WhatsHap (https://github.com/whatshap/whatshap) to extract phase information in the sequence reads and from the subset of available trios and pedigrees, and this information was fed into Shapeit (v.4.2.2; https://odelaneau.github.io/shapeit4) through the --use-PS 0.0001 option. Phasing was carried out in chunks of 10,000 and 100,000 variants (WES and WGS, respectively) and using 500 SNPs from the array data as a buffer at the beginning and end of each chunk. The use of the phased scaffold of array variants meant that chunks of phased sequencing data could be concatenated together to produce whole chromosome files that preserved the chromosome-wide phasing of array variants. A consequence of this process is that when a variant appeared in both the array and sequencing datasets, the data from the array dataset were used.

To assess the performance of the WGS phasing process, we repeated the phasing of chromosome2 by removing the children of the 200 motherfatherchild trios. We then compared the phase of the trio parents to that in the phased dataset that included the children. We observed a mean switch error rate of 0.0024. Without using WhatsHap to leverage phase information in sequencing reads, the mean switch error rate increased to 0.0040 (Supplementary Fig. 23).

The relatedness-inference criteria and relationship assignments were based on kinship coefficients and probability of zero IBD sharing from the KING software (https://www.kingrelatedness.com). We reconstructed all first-degree family networks using PRIMUS (v.1.9.0; https://primus.gs.washington.edu/primusweb) applied to the IBD-based KING estimates of relatedness along with the genetically derived sex and reported age of each individual. In total, 99.3% of the first-degree family networks were unambiguously reconstructed. To visualize the relationship structure in the MCPS, we used the software Graphviz (https://graphviz.org) to construct networks such as those presented in Supplementary Fig. 5. We used the sfdp layout engine which uses a spring model that relies on a force-directed approach to minimize edge length.

To identify IBD segments and to measure ROH, we ran hap-ibd (v.1.0; https://github.com/browning-lab/hap-ibd) using the phased array dataset of 138,511 samples and 538,614 sites from autosomal loci. Hap-ibd was run with the parameter min-seed=4, which looks for IBD segments that are at least 4cM long. We filtered out IBD segments in regions of the genome with fourfold more or fourfold less than the median coverage along each chromosome following the procedure in IBDkin (https://github.com/YingZhou001/IBDkin), and filtered out segments overlapping regions with fourfold less than the median SNP marker density (Supplementary Fig. 28). For the homozygosity analysis, we intersected the sample with the exome data to evaluate loss-of-function variants, which resulted in a sample of 138,200. We further overlaid the ROH segments with local ancestry estimates, and assigned ancestry where the ancestries were concordant between haplotypes and posterior probability was >0.9, assigning ancestry to 99.8% of the ROH.

We used the workflow implemented in the R package bigsnpr (https://privefl.github.io/bigsnpr). In brief, pairwise kinship coefficients were estimated using Plink (v.2.0) and samples were pruned for first-degree and second-degree relatedness (kinship coefficient<0.0884) to obtain a set of unrelated individuals. LD clumping was performed with a default LD r2 threshold of 0.2, and regions with long-range LD were iteratively detected and removed using a procedure based on evaluating robust Mahalanobis distances of PC loadings. Sample outliers were detected using a procedure based on K-nearest neighbours. PC scores and loadings for the first 20 PCs were efficiently estimated using truncated singular value decomposition (SVD) of the scaled genotype matrix. After removal of variant and sample outliers, a final iteration of truncated SVD was performed to obtain the PCA model. The PC scores and loadings from this model were then used to project withheld samples, including related individuals, into the PC space defined by the model using the online augmentation, decomposition and procustes algorithm. For each PC analysis in this study, variants with MAF<0.01 were removed.

Admixture (v.1.3.0; https://dalexander.github.io/admixture) was used to estimate ancestry proportions in a set of 3,964 reference samples representing African, European, East Asian, and American ancestries from a dataset of merged genotypes. This included 765 samples of African ancestry from 1000Genomes (n=661) and HGDP (n=104), 658 samples of European ancestry from 1000Genomes (n=503) and HGDP (n=155), 727 samples of East Asian ancestry from 1000Genomes (n=504) and HGDP (n=223), and 1,814 American samples, including 716 Indigenous Mexican samples from the MAIS study, 64 admixed Mexican American samples from MXL, 21 Maya and 13 Pima samples from HGDP, and 1,000 unrelated Mexican samples from the MCPS. Included SNPs were limited to variants present on the Illumina GSAv.2 genotyping array for which TOPMed-imputed variants in the MAIS study had information r20.9 (m=199,247 SNPs). To select the optimum number of ancestry populations (K) to include in the admixture model, fivefold cross validation was performed for each K in the set 4 to 25 with the cv flag. To obtain ancestry proportion estimates in the remaining set of 137,511 MCPS samples, the population allele frequencies (P) estimated from the analysis of reference samples were fixed as parameters so that the remaining samples could be projected into the admixture model. Projection was performed for the K=4 model and for the K=18 model that produced the lowest cross-validation error, and point estimation was attained using the block relaxation algorithm.

The MAIS genotyping datasets were obtained from L.Orozco from Insituto Nacional de Medicina Genmica. For 644 samples, genotyping was performed using an Affymetrix Human 6.0 array (n=599,727 variants). An additional 72 samples (11 ancestry populations) were genotyped using an Illumina Omni 2.5 array (n=2,397,901 variants). The set of 716 Indigenous samples represent 60 of out the 68 recognized ethnic populations in Mexico3. Per chromosome, VCFs for each genotyping array were uploaded to the TOPMed imputation server (https://imputation.biodatacatalyst.nhlbi.nih.gov) and imputed from a multi-ethnic reference panel of 97,256 whole genomes. Phasing and imputation were performed using the programs eagle and MiniMac, respectively. The observed coefficient of determination (r2) for the reference allele frequency between the reference panel and the genotyping array was 0.696 and 0.606 for the Affymetrix and Illumina arrays, respectively.

Physical positions of imputed variants were mapped from genome build GRCh37 to GRCh38 using the program LiftOver, and only variant positions included on the Affymetrix GSA v.2 were retained. After further filtering out variants with imputation information r2<0.9, the following QC steps were performed before merging of the MAIS Affymetrix and Illumina datasets: (1) removal of ambiguous variants (that is, A/T and C/G polymorphisms); (2) removal of duplicate variants; (3) identifying and correcting allele flips; and (4) removal of variants with position mismatches. Merging was performed using the --bmerge command in Plink (v.1.9).

We used publicly available genotypes from the HGDP (n=929) and the 1000Genomes project (n=2,504). To obtain a combined global reference dataset for downstream analyses of population structure, admixture and local ancestry, the HGDP and 1000Genomes datasets were merged. The resulting merged public reference dataset was subsequently merged with the MAIS dataset and MCPS genotyping array dataset. Each merge was performed using the bmerge function in Plink (v.1.9; https://www.cog-genomics.org/plink) after removing ambiguous variants, removing duplicate variants, identifying and correcting allele flips, and removing variants with position mismatches. The combined global reference dataset comprised 199,247 variants and 142,660 samples.

To characterize genetic admixture within the MCPS cohort, we performed a seven-way LAI analysis with RFMix (v.2.0; https://github.com/slowkoni/rfmix) that included reference samples from the HGDP and 1000Genomes studies, and Indigenous samples from the MAIS study. This merged genotyping dataset of samples across these studies with the 138,511 MCPS participants included 204,626 autosomal variants and 5,363 chromosomeX variants.

To identify reference samples with extensive admixture to exclude from LAI, we performed admixture analysis with the program TeraSTRUCTURE (https://github.com/StoreyLab/terastructure) on a merged genotyping dataset (n=3,274) that included African (AFR), European (EUR) and American (AMR) samples from the HGDP, 1000Genomes and MAIS studies, and 1,000 randomly selected unrelated MCPS samples. Following the recommended workflow in the TeraSTRUCTURE documentation (https://github.com/StoreyLab/terastructure), we varied the rfreq parameter from the set of {0.05, 0.10, 0.15, 0.20} of autosomal variants with K=4 and selected the value that maximized the validation likelihood (20% of autosomal variants; rfreq=45,365). We then varied the K parameter and ran it in triplicate to identify the value that attained a maximal average validation likelihood (K=18). Each of the estimated K ancestries was assigned to a global superpopulation (that is, AFR, EUR and AMR), and the cumulative K ancestry proportion was used as an ancestry score for selecting reference samples. Using an ancestry score threshold of 0.9, 666 AFR, 659 EUR and 616 AMR samples were selected as reference samples. The AMR samples used for seven-way LAI comprised 98 Mexico_North, 42 Mexico_Northwest, 185 Mexico_Central, 128 Mexico_South and 163 Mexico_Southeast individuals.

Reference samples were phased using Shapeit (v.4.1.2; https://odelaneau.github.io/shapeit4) with default settings, and the phasing of the 138,511 MCPS participants was performed as described above (see the section Array phasing). Seven-way LAI was performed using RFMix (v.2.0), with the number of terminal nodes for the random forest classifier set to 5 (-n 5), the average number of generations since expected admixture set to 15 (-G 15), and ten rounds of expectation maximization (EM) algorithm (-e 10). Global ancestry proportion estimates were derived by taking the average per-chromosome Q estimates (weighted by chromosome length) for each of the seven ancestries (that is, AFR, EUR, Mexico_North, Mexico_Northwest, Mexico_Central, Mexico_South and Mexico_Southeast). Inferred three-way global ancestry proportion estimates were obtained by combining proportions for each of the five Indigenous Mexican populations into a single AMR category.

To delineate local ancestry segments for use in the estimation of ancestry-specific allele frequencies (see the section Ancestry-specific allele frequency estimation), we performed a three-way LAI analysis using a merged genotyping dataset that excluded the MAIS samples as this afforded greater genotyping density (493,036 autosomal variants and 12,798 chromosomeX variants). Before LAI analysis, reference samples were selected using the same workflow for TeraSTRUCTURE as described above, with modifications being the inclusion of 10,000 unrelated MCPS participants and an ancestry threshold of 0.95. RFMix was applied as described above, with modifications being the use of 753 AFR, 649 EUR and 91 AMR reference samples, specification of 5 rounds of EM (-e 5), and use of the --reanalyze-reference option, which treated reference haplotypes as if they were query haplotypes and updated the set of reference haplotypes in each EM round.

To measure the correlation in ancestry between partner pairs, we used a linear model to predict ancestry of each partner using the ancestry of their spouse, education level (four categories) and district (Coyoacn and Iztapalapa) of both partners.

We averaged local ancestry dosages (estimated using RFMix at 98,012 positions along the genome) from 78,833 unrelated MCPS samples and performed a per-ancestry scan testing for deviation of local ancestry proportion from the global ancestry proportion19. The test is based on assumptions of binomial sampling and normal approximation for the sample mean. The global ancestry proportion for each ancestry was estimated as a robust average over local ancestry using the Tukeys biweight robust mean. The scan was performed in all autosomes separately for African, European and Indigenous Mexican ancestries with the significance threshold 1.7107=0.05/(98, 0123), which accounts for the number of local ancestry proportions tested and the three ancestries.

IBD segments from hapIBD were summed across pairs of individuals to create a network of IBD sharing represented by the weight matrix (Win {{mathbb{R}}}_{ge 0}^{ntimes n}) for n samples. Each entry ({w}_{{ij}}in W) gives the total length in cM of the genome that individuals i and j share identical by descent. We sought to create a low-dimensional visualization of the IBD network. We used a similar approach to that described in ref. 14, which used the eigenvectors of the normalized graph Laplacian as coordinates for a low-dimensional embedding of the IBD network. Let D be the degree matrix of the graph with ({d}_{{ii}}=sum _{{j}}{w}_{{ij}}) and 0 elsewhere. The normalized (random walk) graph Laplacian is defined to be (L=I-{D}^{-1}W), where I is the identity matrix.

The matrix L is positive semi-definite, with eigenvalues (0={lambda }_{0}le {lambda }_{1}le cdots le {lambda }_{n-1}). The multiplicity of eigenvalue 0 is determined by the number of connected components in the IBD network. If L is fully connected, the eigenvector associated with eigenvalue 0 is constant, whereas the remaining eigenvectors can be used to compute a low-dimensional representation of the IBD network. If p is the desired dimension, and u1,, up the bottom 1p eigenvectors of L (indexed from 0), the matrix (Uin {{mathbb{R}}}^{ntimes p}) with columns u1,, up define a low-dimensional representation of each individual in the IBD network39. In practice, we solved the generalized eigenvalue problem to obtain u1,, up.

If u is an eigenvector of L with eigenvalue , then u solves the generalized eigenvalue problem with eigenvalue 1.

To apply to the IBD network of the MCPS cohort, we first removed edges with weight >72cM as previously done14. We did this to avoid the influence on extended families on the visualization. We next extracted the largest connected component from the IBD network, and computed the bottom u1,, u20 eigenvectors of the normalized graph Laplacian.

To examine fine-scale population structure using haplotype sharing, we calculated a haplotype copying matrix L using Impute5 (https://jmarchini.org/software/#impute-5) with entries Lij that are the length of sequence individual i copies from individual j. Impute5 uses a scalable imputation method that can handle very large haplotype reference panels. At its core is an efficient Hidden Markov model that can estimate the local haplotype sharing profile of a target haplotype with respect to a reference set of haplotypes. To avoid the costly computations of using all the reference haplotypes, an approach based on the PBWT data structure was used to identify a subset of reference haplotypes that led to negligible loss of accuracy. We leveraged this methodology to calculate the copying matrix L, using array haplotypes from a set of 58,329 unrelated individuals as both target and reference datasets, and used the --ohapcopy ban-repeated-sample-names flags to ban each target haplotype being able to copy itself. SVD on a scaled centred matrix was performed using the bigstatsr package (https://cran.r-project.org/web/packages/bigstatsr/index.html) to generate 20 PCs. This is equivalent to an eigen-decomposition of the variance-covariance matrix of recipients shared segment lengths.

We imputed the filtered array dataset using both the MCPS10k reference panel and the TOPMed imputation server. For TOPMed imputation, we used Plink2 to convert this dataset from Plink1.9 format genotypes to unphased VCF genotypes. For compatibility with TOPMed imputation server restrictions, we split the samples in this dataset into six randomly assigned subsets of about 23,471 samples, and into chromosome-specific bgzipped VCF files. Using the NIH Biocatalyst API (https://imputation.biodatacatalyst.nhlbi.nih.gov), we submitted these six jobs to the TOPMed imputation server. Following completion of all jobs, we used bcftools merge to join the resulting dosage VCFs spanning all samples. For the MCPS10k imputation, we used Impute5 (v.1.1.5). Each chromosome was split into chunks using the imp5Chunker program with a minimum window size of 5Mb and a minimum buffer size of 500kb. Information scores were calculated using qctool (https://www.well.ox.ac.uk/~gav/qctool_v2/).

The 1000Genomes WGS genotype VCF files were downloaded (http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data_collections/1000G_2504_high_coverage/working/20201028_3202_phased/) and filtered to remove sites that are multi-allelic sites, duplicated, have missingness >2%, HardyWeinberg P<1108 in any subpopulation and MAF<0.1% in any subpopulation. We used only those 490 AMR samples in the MXL, CLM, PUR and PEL subpopulations. We constructed two subsets of genotypes on chromosome2 from the Illumina HumanOmniExpressExome (8.v1-2) and Illumina GSA (v.2) arrays, and these were used as input to the TOPMed and MCPS10k imputation pipelines.

We measured imputation accuracy by comparing the imputed dosage genotypes to the true (masked) genotypes at variants not on the arrays. Markers were binned according to the MAF of the marker in 490 AMR samples. In each bin, we report the squared correlation (r2) between the concatenated vector of all the true (masked) genotypes at markers and the vector of all imputed dosages at the same markers. Variants that had a missing rate of 100% in the WGS dataset before phasing were removed from the imputation assessment.

The LAI results consist of segments of inferred ancestry across each haplotype of the phased array dataset. As the WES and WGS alleles were phased onto the phased array scaffold, we inferred the ancestry of each exome allele using interpolation from the ancestry of the flanking array sites. For each WES and WGS variant on each phased haplotype, we determined the RFMix ancestry probability estimates at the two flanking array sites and used their relative base-pair positions to linearly interpolate their ancestry probabilities. For a given site, if ({p}_{{ijk}}) is the probability that the jth allele of the ith individual is from population k, and Gij is the 0/1 indicator of the non-reference allele for the jth allele of the ith individual then the weighted allele count (ACk), the weight allele number (ANk) and the allele frequency (k) of the kth population is given by

$${{rm{AC}}}_{k}=mathop{sum }limits_{i=1}^{n}mathop{sum }limits_{j=1}^{2}{p}_{ijk}{G}_{ij},,{{rm{AN}}}_{k}=mathop{sum }limits_{i=1}^{n}mathop{sum }limits_{j=1}^{2}{p}_{ijk},,{theta }_{k}=frac{{{rm{AC}}}_{k}}{{{rm{AN}}}_{k}}$$

An estimate of the effective sample size for population k at the site is ({n}_{k}={{rm{AN}}}_{k}/2). Singleton sites can be hard to phase using existing methods. Family information and phase information in sequencing reads was used in the WGS phasing, and this helped to phase a proportion of the singleton sites. In the WES dataset, we found that 46% of exome singletons occurred in stretches of heterozygous ancestry. For these variants, we gave equal weight to the two ancestries when estimating allele frequencies.

To validate the MCPS allele frequencies, we downloaded the gnomAD v.3.1 reference dataset (https://gnomad.broadinstitute.org) and retained only high-quality variants annotated as passed QC (FILTER=PASS), SNVs, outside low-complexity regions and with the number of called samples greater than 50% of the total sample size (n=76,156). We additionally overlapped gnomAD variants with TOPMed Freeze8 high-quality variants (FILTER=PASS) (https://bravo.sph.umich.edu/freeze8/hg38). We further merged gnomAD variants and MCPS exome variants by the chromosome, position, reference allele and alternative allele names and excluded MCPS singletons, which were heterozygous in ancestry. This process resulted in 2,249,986 overlapping variants available for comparison with the MCPS WES data. Median sample sizes in gnomAD non-Finish Europeans, African/Admixed African and Admixed American populations were 34,014, 20,719 and 7,639, respectively.

To investigate the effect of relatedness on allele frequency estimates, we implemented a method to compute relatedness-corrected allele frequencies using identical-by-descent (IBD) segments. This method computes allele frequencies at a locus by clustering alleles inherited IBD from a common ancestor, then counting alleles once per common ancestor rather than once per sample. Because IBD sharing is affected by both demography and relatedness, we limited IBD sharing to segments between third-degree relatives or closer. Conceptually, this is equivalent to tracing the genealogy of a locus back in time across all samples until no third-degree relatives remain, then computing allele frequencies in the ancestral sample.

We estimated allele frequencies in two steps. First, we constructed a graph based on IBD sharing at a locus. Second, we estimated allele counts and allele numbers by counting the connected components of the IBD graph. Our approach is similar to the DASH haplotype clustering approach40. However, we make different assumptions about how errors affect the IBD graph and additionally compute ancestry-specific frequencies using local ancestry inference estimates.

To construct the IBD graph, suppose we have genotyped and phased N diploid samples at L biallelic loci. For each locus l we construct an undirected graph Gl=(Vl,El) describing IBD sharing among haplotypes. Let the tuple (i, j)l represent haplotype j of sample i at locus l, and let ({h}^{{left(i,jright)}_{l}}in {mathrm{0,1}}) be the allele itself. Define

$$begin{array}{l}{V}_{l},=,{{(i,j)}_{l}:{rm{for}},1le jle 2,{rm{and}},1le ile N}\ {E}_{l},=,{({(i,j)}_{l},{(s,t)}_{l}):{h}^{{(i,j)}_{l}},{rm{and}},{h}^{{(s,t)}_{l}},{rm{are}},{rm{IBD}}}.end{array}$$

In words, the set of vertices V constitute all haplotypes at locus l. Each edge in E is between a pair of haplotypes that fall on the same IBD segment (Supplementary Fig. 25).

If IBD segments are observed without error, then each maximal clique of Gl represents a set of haplotypes descended from a common ancestor. In practice, edges will be missing owing to errors in IBD calling. Thus, what we observe are sets of connected components rather than maximal cliques. Because we limited edges to pairs of third-degree relatives or closer, we assumed missing edges in connected components are false negatives and included them. We additionally removed edges between haplotypes for which the observed alleles conflicted.

Given an IBD graph Gl=(Vl, El) for a locus l, we estimated alternative allele counts and allele numbers by counting the connected components of the graph. Let Cl1,,Clm be the connected components of Gl. Let CALT={Cim: haplotypes in Cim have the ALT allele} and CREF={Cim: haplotypes in Cim have the REF allele}

Then

$$begin{array}{l}AC=| {C}_{{rm{ALT}}}| \ AN=| {C}_{{rm{ALT}}}| +| {C}_{{rm{REF}}}| \ AF=AC,/,ANend{array}$$

We additionally used LAI estimates to compute ancestry-specific frequencies. Let ({p}^{{(i,j)}_{l}}in {{mathbb{R}}}^{K}) be the vector of probabilities that an allele on haplotype j from sample i at locus l comes from one of K populations. For each connected component, we averaged local ancestry estimates

$${bar{p}}_{{C}_{im}}=frac{1}{|{C}_{lm}|}{sum }_{{(i,j)}_{l}in {C}_{lm}}{p}^{{(i,j)}_{l}}$$

We computed a vector of weighted allele counts W and allele numbers N by

$$begin{array}{l}W={sum }_{Cin {C}_{{rm{ALT}}}}{bar{p}}_{C}\ N={sum }_{Cin {C}_{{rm{ALT}}}}{bar{p}}_{C}+{sum }_{Cin {C}_{{rm{REF}}}}{bar{p}}_{C}end{array}$$

Ancestry-specific frequencies were estimated by dividing each component of W by the corresponding component of N.

For singletons for which the phasing of haplotypes was unknown, we averaged local ancestry estimates from haplotypes in the sample.

To generate source datasets for assessing trans-ancestry portability of BMI PRS, whole genome regression was performed using Regenie (https://rgcgithub.github.io/regenie/) in individuals in the MCPS and in a predominantly European-ancestry cohort from the UK Biobank. Individuals with type2 diabetes (ICD10 code E11 or self-reported) were excluded. BMI values underwent rank-based inverse normal transformation (RINT) by sex and ancestry; models were additionally adjusted for age, age2 and technical covariates (UK Biobank). The Regenie summary statistics from the UK Biobank were used to generate a BMI PRS in MCPS; conversely, MCPS summary statistics were applied to UK Biobank statistics.

To avoid overfitting with respect to selection of a PRS algorithm and its associated tuning parameters, LDpred (https://github.com/bvilhjal/ldpred) with value of 1 was chosen from a recent publication of BMI and obesity27. Summary statistics were restricted to HapMap3 variants and followed existing filtering recommendations. In the MCPS, two PRS values were generated; imputed variants were obtained from the MCPS10k reference panel or the TOPMed panel. In the UK Biobank data, PRS values were calculated separately by continental ancestry (African, East Asian, European, Latino, South Asian), determined from a likelihood-based inference approach8 in a merged dataset of variants from UK Biobank and the 1000Genomes project.

To evaluate PRS performance, BMI values were transformed (RINT) by sex and ancestry and regressed on PRS, age and age2. As for the generation of summary statistics, individuals with diabetes were excluded from the analysis. PRS accuracy was assessed by incrementalR2 (proportional reduction in regression sum of squares error between models with and without BMI PRS). Additionally, raw BMI values with PRS, age, age2, sex and ancestry were modelled to obtain per BMI PRS standard deviation effect-size estimates. The impact of ancestry differences on source summary statistics compared to target PRS was assessed with two approaches. For the MCPS, individuals were divided into quantiles by estimated Indigenous Mexican Ancestry using the LAI approach described above. For the UK Biobank, metrics were calculated within each 1000Genomes-based continental ancestry.

The MCPS represents a long-standing scientific collaboration between researchers at the National Autonomous University of Mexico and the University of Oxford, who jointly established the study in the mid-1990s and have worked together on it ever since. Blood sample collection and processing were funded by a Wellcome Trust grant to the Mexican and Oxford investigators. However, at the time, no funding was requested to create an appropriate long-term sample storage facility in Mexico City. Therefore, the Mexican investigators agreed for the samples to be shipped to Oxford where they could be stored in a liquid-nitrogen sample storage archive (funded by the UK Medical Research Council and Cancer Research UK) that had previously been established by the Oxford team, and only on the understanding that control of the samples remained with the Mexican investigators. The shipping of blood samples from Mexico to the United Kingdom was approved by the Mexican Ministry of Health, and the study was approved by scientific and ethics committees within the Mexican National Council of Science and Technology (0595 P-M), the Mexican Ministry of Health and the Central Oxford Research Ethics Committee (C99.260). Although appropriate facilities in Mexico City now exist to store the samples, the Mexican investigators have decided that the costs of sending them back to Mexico exceed the benefits of having closer access to them. Study participants gave signed consent in keeping with accepted ethical practices at the time for observational cohort studies. The baseline consent form stated that their blood samples would be stored and used in the future for unspecified research purposes (with a specific statement that this would include future analysis of genetic factors) and that it would probably be many years before such blood analyses were done. The MCPS consent form also stated that the research was being done in collaboration with the University of Oxford and that the purpose of the study was to benefit future generations of Mexican adults. In 2019, the Mexican and Oxford investigators jointly agreed to allow the extracted DNA to be sent to the Regeneron Genetics Center after they had offered to genotype and exome sequence the entire cohortthereby creating the resource now available for future research by Mexican scientists (see the Data Availability section)in exchange for sharing the other data with them for the purpose of performing joint collaborative genetic analyses. Formal approval to share MCPS data with commercial institutions was sought and obtained from the Medical Ethics Committee of the National Autonomous University of Mexico (FMED/CEI/MHU/001/2020). Major discoveries from the study have been disseminated through open-access scientific publications, local and international scientific meetings, press releases, social media and local television, but direct communication of study results to the original study participants is unfortunately not practical as no information on telephone numbers or email addresses was collected at recruitment. As in other prospective cohort studies (such as the UK Biobank), it was agreed that there would be no feedback of individual blood results to participants, as it has been shown that such feedback can do more harm than good (whereas no feedback ensures that that is not the case).

Recruitment of individuals in the MAIS cohort was done with approval of the leaders of the Indigenous communities and with the support of the National Commission for the Development of Indigenous Communities of Mexico (CDI), now the Instituto Nacional de los Pueblos Indgenas (INPI). All participants provided written informed consent, and authorities or community leaders participated as translators where necessary. The consent form described how findings from the study may have commercial value and be used by for-profit companies. Sample collection for MAIS was approved by the Bioethics and Research Committees of the Insituto Nacional de Medicina Genmica in Mexico City (protocol numbers 31/2011/I and 12/2018/I). Preliminary data from the MAIS cohort have been discussed with the Indigenous leaders and volunteer individuals included in the study, explaining the meaning of the findings on health or populations history, and the potential use of the data in future collaborations.

Further information on research design is available in theNature Portfolio Reporting Summary linked to this article.

Link:
Genotyping, sequencing and analysis of 140,000 adults from Mexico ... - Nature.com

Domcke, S. et al. Evaluating cell lines as tumour models by comparison of genomic profiles. Nat. Commun. 4, 2126 (2013).

Article ADS PubMed Google Scholar

Kbel, M. et al. Ovarian carcinoma subtypes are different diseases: Implications for biomarker studies. PLoS Med. 5(12), e232 (2008).

Article PubMed PubMed Central Google Scholar

Soslow, R. A. Histologic subtypes of ovarian carcinoma: An overview. Int. J. Gynecol. Pathol. 27(2), 161172 (2008).

PubMed Google Scholar

Ip, C. K. M. et al. Stemness and chemoresistance in epithelial ovarian carcinoma cells under shear stress. Sci. Rep. 6, 26788 (2016).

Article ADS CAS PubMed PubMed Central Google Scholar

Galluzzi, L. et al. Systems biology of cisplatin resistance: Past, present and future. Cell Death Dis. 5(5), e1257e1257 (2014).

Article CAS PubMed PubMed Central Google Scholar

Seborova, K. et al. Association of ABC gene profiles with time to progression and resistance in ovarian cancer revealed by bioinformatics analyses. Cancer Med. 8(2), 606616 (2019).

Article CAS PubMed PubMed Central Google Scholar

Sharom, F. J. ABC multidrug transporters: structure, function and role in chemoresistance. Pharmacogenomics 9(1), 105127 (2008).

Article CAS PubMed Google Scholar

Smida, M. & Nijman, S. Functional drug-gene interactions in lung cancer. Expert Rev. Mol. Diagn. 12(3), 291302 (2012).

Article CAS PubMed Google Scholar

Shanker, S. et al. Drug resistance in lung cancer. Lung Cancer: Targets Ther. 1, 2336 (2010).

CAS Google Scholar

Chung, F. S. et al. Disrupting P-glycoprotein function in clinical settings: What can we learn from the fundamental aspects of this transporter?. Am. J. Cancer Res. 6(8), 15831598 (2016).

CAS PubMed PubMed Central Google Scholar

Derti, A. et al. A quantitative atlas of polyadenylation in five mammals. Genome Res. 22(6), 11731183 (2012).

Article CAS PubMed PubMed Central Google Scholar

Shi, Y. Alternative polyadenylation: New insights from global analyses. RNA 18(12), 21052117 (2012).

Article CAS PubMed PubMed Central Google Scholar

Reyes, A. & Huber, W. Alternative start and termination sites of transcription drive most transcript isoform differences across human tissues. Nucleic Acids Res. 46(2), 582592 (2017).

Article PubMed Central Google Scholar

Mandel, C., Bai, Y. & Tong, L. Protein factors in pre-mRNA 3-end processing. Cell. Mol. Life Sci. 65(78), 10991122 (2008).

Article CAS PubMed PubMed Central Google Scholar

Danckwardt, S., Hentze, M. W. & Kulozik, A. E. 3 end mRNA processing: Molecular mechanisms and implications for health and disease. EMBO J. 27(3), 482498 (2008).

Article CAS PubMed PubMed Central Google Scholar

Neilson, J. R. & Sandberg, R. Heterogeneity in mammalian RNA 3 end formation. Exp. Cell Res. 316(8), 13571364 (2010).

Article CAS PubMed PubMed Central Google Scholar

Chekulaeva, M., Hentze, M. W. & Ephrussi, A. Bruno acts as a dual repressor of oskar translation, promoting mRNA oligomerization and formation of silencing particles. Cell 124(3), 521533 (2006).

Article CAS PubMed Google Scholar

Berkovits, B. D. & Mayr, C. Alternative 3 UTRs act as scaffolds to regulate membrane protein localization. Nature 522(7556), 363367 (2015).

Article ADS CAS PubMed PubMed Central Google Scholar

An, J. J. et al. Distinct role of long 3 UTR BDNF mRNA in spine morphology and synaptic plasticity in hippocampal neurons. Cell 134(1), 175187 (2008).

Article CAS PubMed PubMed Central Google Scholar

Gruber, A. R. et al. Global 3 UTR shortening has a limited effect on protein abundance in proliferating T cells. Nat. Commun. 5, 5465 (2014).

Article ADS CAS PubMed Google Scholar

Magee, P., Shi, L. & Garofalo, M. Role of microRNAs in chemoresistance. Ann. Transl. Med. 3(21), 332 (2015).

PubMed PubMed Central Google Scholar

Tanaka, I. Regulation of alternative splicing and polyadenylation by genotoxic anti-cancer agents. Universit Paris-Saclay (2019).

Cole, S. P. C. Targeting multidrug resistance protein 1 (MRP1, ABCC1): Past, present, and future. Annu. Rev. Pharmacol. Toxicol. 54(1), 95117 (2014).

Article CAS PubMed Google Scholar

Hipfner, D. R. et al. Monoclonal antibodies that inhibit the transport function of the 190-kDa multidrug resistance protein, MRP. Localization of their epitopes to the nucleotide-binding domains of the protein. J. Biol. Chem. 274(22), 1542015426 (1999).

Article CAS PubMed Google Scholar

Emmanouilidi, A. et al. Inhibition of the lysophosphatidylinositol transporter ABCC1 reduces prostate cancer cell growth and sensitizes to chemotherapy. Cancers (Basel) 12(8), 2022 (2020).

Article CAS PubMed Google Scholar

Fanelli, M. et al. Targeting ABCB1 and ABCC1 with their specific inhibitor CBT-1 can overcome drug resistance in osteosarcoma. Curr. Cancer Drug Targets 16(3), 261274 (2016).

Article MathSciNet CAS PubMed Google Scholar

Stefan, S. M. Multi-target ABC transporter modulators: What next and where to go?. Future Med. Chem. 11(18), 23532358 (2019).

Article CAS PubMed Google Scholar

Schumacher, T. et al. ABC transporters B1, C1 and G2 differentially regulate neuroregeneration in mice. PLOS ONE 7(4), e35613 (2012).

Article ADS CAS PubMed PubMed Central Google Scholar

Galluzzi, L. et al. Molecular mechanisms of cell death: Recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death Differ. 25(3), 486541 (2018).

Article PubMed PubMed Central Google Scholar

Basuli, D. et al. Iron addiction: A novel therapeutic target in ovarian cancer. Oncogene 36(29), 40894099 (2017).

Article CAS PubMed PubMed Central Google Scholar

Li, D., Zhang, M. & Chao, H. Significance of glutathione peroxidase 4 and intracellular iron level in ovarian cancer cellsutilization of ferroptosis mechanism. Inflamm. Res. 70(10), 11771189 (2021).

Article CAS PubMed Google Scholar

Chen, L. et al. Erastin sensitizes glioblastoma cells to temozolomide by restraining xCT and cystathionine--lyase function. Oncol. Rep. 33(3), 14651474 (2015).

Article MathSciNet CAS PubMed Google Scholar

Yamaguchi, H. et al. Caspase-independent cell death is involved in the negative effect of EGF receptor inhibitors on cisplatin in non-small cell lung cancer cells. Clin. Cancer Res: Off. J. Am. Assoc. Cancer Res. 19(4), 845854 (2013).

Article CAS Google Scholar

Zhou, H.-H. et al. Erastin reverses ABCB1-mediated docetaxel resistance in ovarian cancer. Front. Oncol. 9, 13981398 (2019).

Article PubMed PubMed Central Google Scholar

Beaufort, C. M. et al. Ovarian cancer cell line panel (OCCP): Clinical importance of in vitro morphological subtypes. PLOS ONE 9(9), e103988 (2014).

Article ADS PubMed PubMed Central Google Scholar

Perets, R. et al. Transformation of the fallopian tube secretory epithelium leads to high-grade serous ovarian cancer in Brca;Tp53;Pten models. Cancer cell 24(6), 751765 (2013).

Article CAS PubMed PubMed Central Google Scholar

Klotz, D. M. & Wimberger, P. Cells of origin of ovarian cancer: Ovarian surface epithelium or fallopian tube?. Arch. Gynecol. Obstet. 296(6), 10551062 (2017).

Article PubMed Google Scholar

Soong, T. R. et al. The fallopian tube, precursor escape and narrowing the knowledge gap to the origins of high-grade serous carcinoma. Gynecol. Oncol. 152(2), 426433 (2019).

Article PubMed Google Scholar

Ducie, J. et al. Molecular analysis of high-grade serous ovarian carcinoma with and without associated serous tubal intra-epithelial carcinoma. Nat. Commun. 8(1), 990 (2017).

Article ADS PubMed PubMed Central Google Scholar

Masamha, C. P. & Todd, Z. Adapting 3 rapid amplification of CDNA ends to map transcripts in cancer. J. Vis. Exp. 133, e57318 (2018).

Google Scholar

Bruhn, O. et al. Alternative polyadenylation of ABC-transporters of the C-family (ABCC1, ABCC2, ABCC3) and implications on post-transcriptional micro-RNA regulation. Mol. Pharmacol. 97, 112122 (2019).

Article PubMed Google Scholar

Kurnit, K. C., Fleming, G. F. & Lengyel, E. Updates and new options in advanced epithelial ovarian cancer treatment. Obstet. Gynecol. 137(1), 108 (2021).

Article CAS PubMed Google Scholar

Gao, B. et al. Paclitaxel sensitivity in relation to ABCB1 expression, efflux and single nucleotide polymorphisms in ovarian cancer. Sci. Rep. 4, 4669 (2014).

Article CAS PubMed PubMed Central Google Scholar

Mayr, C. Regulation by 3-untranslated regions. Annu. Rev. Genetics 51(1), 171194 (2017).

Article CAS Google Scholar

Friedman, R. C. et al. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res. 19(1), 92105 (2009).

Article CAS PubMed PubMed Central Google Scholar

Mayr, C. & Bartel, D. P. Widespread shortening of 3UTRs by alternative cleavage and polyadenylation activates oncogenes in cancer cells. Cell 138(4), 673684 (2009).

Article CAS PubMed PubMed Central Google Scholar

Masamha, C. P. et al. CFIm25 links alternative polyadenylation to glioblastoma tumour suppression. Nature 510(7505), 412416 (2014).

Article ADS CAS PubMed PubMed Central Google Scholar

He, X.-J. et al. Aberrant alternative polyadenylation is responsible for survivin up-regulation in ovarian cancer. Chin. Med. J. 129(10), 11401146 (2016).

Article CAS PubMed PubMed Central Google Scholar

Patch, A.-M. et al. Wholegenome characterization of chemoresistant ovarian cancer. Nature 521(7553), 489494 (2015).

Article ADS CAS PubMed Google Scholar

Chen, M. et al. miR-133b down-regulates ABCC1 and enhances the sensitivity of CRC to anti-tumor drugs. Oncotarget 8(32), 5298352994 (2017).

Article PubMed PubMed Central Google Scholar

Ma, J. et al. Involvement of miR-133a and miR-326 in ADM resistance of HepG2 through modulating expression of ABCC1. J. Drug Target 23(6), 519524 (2015).

More:
The role and impact of alternative polyadenylation and miRNA ... - Nature.com