Jack Kosmicki
Statistical geneticist @ Regeneron
tweets are my own | he/him/his
- Reposted by Jack KosmickiToday in @nature.com, we describe how discarded reads in biobank-scale WGS can help resolve the genetic predictors and consequences of Epstein-Barr Virus (EBV) infection. Wonderful working with @ryandhindsa.bsky.social @sherrynyeo.bsky.social @erinmayc.bsky.social www.nature.com/articles/s41...
- Reposted by Jack KosmickiA new preprint from Peter Mchale and Michael Goldberg in my group on the latent biases inherent to current models of non-coding constraint. www.biorxiv.org/content/10.6...
- It's truly a sad state of affairs when it takes longer to pull GWAS sumstats out of the AllofUs research platform than it takes to generate them.
- Really nice work by @nbaya.bsky.social and co. showing that, as expected, individuals whose observed phenotype deviates from their genetically predicted trait are enriched for rare damaging variants in associated genes with said trait. Now to see if embryo selection companies pick up on this 😉
- Why do some individuals defy their polygenic score? In the largest study of its kind (402k UKB individuals; 7 continuous traits + 3 diseases), we asked: If your phenotype deviates from common-variant polygenic score prediction, what's driving that difference? www.medrxiv.org/content/10.6...
- Reposted by Jack KosmickiI wrote about the bizarre case of Herasight, the embryo selection company going all in on eugenics.
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack KosmickiAnother #aDNA preprint! We recovered DNA from 7,000 year old (!) goat leather from Cueva de los Murciélagos - and see a genetic link with Bermeya goats today! Credit to Francisco Martínez-Sevilla for seeing me on Youtube(?!) + reaching out to collaborate - there's still value in the internet.
- Reposted by Jack KosmickiExcited to share our new FinnGen single-nucleus multiome preprint! 🧬 We profiled ~10M PBMCs (snRNA-seq + snATAC-seq) from 1,108 Finnish donors to map how genetic variants drive complex disease through chromatin and gene regulation 🧵👇 🔗 Link: www.medrxiv.org/content/10.1...
- Reposted by Jack KosmickiMassive single-cell study by Kanai et al (www.medrxiv.org/content/10.1...): - Once statistical power is high, constrained genes have more (though weaker) eQTLs. - Chromatin-QTLs near constrained genes have "normal" effect sizes, colocalize more with disease, but exhibit attenuated peak-gene effects.
- 🚨Nearly all GWAS is done in adults - making inferences about genetic effects in children or child-specific traits/diseases impossible to know. A huge study of 80.6k Japanese children covering 1.1k GWASes (many for the first time) by is now out www.medrxiv.org/content/10.1...
- Great use of AI 🙄 www.nature.com/articles/s41...
- Yes, Boston's food scene is . . . underwhelming at best. The only restaurant to receive a ⭐️ was 311 Omakase - not the local favorites - Sarma, Oleana, Mooncusser, Nightshade Noodle Bar (which frankly are all massively overrated). www.masslive.com/boston/2025/...
- Reposted by Jack KosmickiOur paper on imputation of ancient goat genomes is now available at GBE - congratulations to @jolijnerven.bsky.social #aDNA
- 🚨I did my PhD at Harvard and it fundamentally shaped who I am today. It's extremely sad to see their PhD admission cuts for the next 2 years. - Science ⬇️75% - Arts & Humanities ⬇️60% - Social Sciences ⬇️50–70% - History ⬇️60% - Org. & Evo. Bio. ⬇️75% - German ⬇️100% www.thecrimson.com/article/2025...
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki📃 We’re excited to share our latest work, now published in Nature Communications — a major update to the Genome Aggregation Database (gnomAD) that improves allele frequency resolution for two gnomAD-defined genetic ancestry groups using local ancestry inference (LAI).
- Reposted by Jack KosmickiVery happy to see our pre-print on ancient Irish goat genetics on bioRxiv #aDNA #spiergorm I want to acknowledge this was only possible through the work of the late Dr. Judith Findlater, along with Prof. Eileen Murphy at @qubelfastofficial.bsky.social.
- Reposted by Jack KosmickiInterested in goats 🐐? Read our pre-print on ancient Irish goats and their connection to the Old Irish goat breed Just look at this dashing individual, courtesy of the Old Irish Goat society
- Very happy to see our pre-print on ancient Irish goat genetics on bioRxiv #aDNA #spiergorm I want to acknowledge this was only possible through the work of the late Dr. Judith Findlater, along with Prof. Eileen Murphy at @qubelfastofficial.bsky.social.
- Reposted by Jack Kosmicki[Not loaded yet]
- Kristin Tsuo, @genetisaur.bsky.social, & Mark Daly just wrote the best proteomics paper I've read. They convincingly show how smoking and alcohol (aka, the environment) dramatically influences proteomics data. Some 🤯 results 1) proteomics predicts frequency (5a) & quanity (5b) of alcohol consumed
- 2) proteomics stratifies smokers, prior smokers, & non smokers (Fig. 4a) 3) proteomics even stratifies how LONG it has been since you quit smoking (Fig. 4b)🤯
- Reposted by Jack KosmickiI will be recruiting two PhDs for my ERC Project HERDPATH - let's discover out how livestock and pathogens evolved together using #aDNA. Projects will be animal or pathogen focused but will be in dialogue. Details at my quaint website below. Deadline 3rd October. kevingdaly.github.io
- Reposted by Jack Kosmicki[Not loaded yet]
- 😂 A WashU professor embezzled $412k and used the money to buy . . . *checks notes* . . . collectible trading cards www.justice.gov/usao-edmo/pr...

- Reposted by Jack Kosmicki
- Reposted by Jack Kosmicki[Not loaded yet]
- A tale of two DTC DNA companies. While 23andMe is dealing with the bankruptcy fallout, Ancestry is sponsoring Ryan Reynolds' and Rob McElhenney's Welsh football club, Wrexham AFC. PS: I loved watching Welcome to Wrexham ❤️ www.wrexhamafc.co.uk/news/2025/ju...
- In the latest update to the 23andMe, Anne Wojcicki (i.e., not REGN) won the bidding war. www.wsj.com/tech/biotech...
- Year ago, the prevailing attitude was that deep learning would significantly improve polygenic risk scores (or anything else, to be fair). A rather important set of negative results indicating that isn't the case (though I imagine many will still try). nature.com/articles/s41...
- Whelp, the sale was premature. Former CEO, Anne Wojcicki, reopened the 23andMe auction with a new $305 million bid. www.bloomberg.com/news/article...
- Wow, Regeneron bought 23andMe for $256 million. www.cnn.com/2025/05/19/b...
- Reposted by Jack KosmickiThere are 2 previous historical cases of countries destroying their science and universities, crippling them for decades: Lysenkoism in the USSR and Nazi Germany. The Trump administration will be the 3rd. It's not just budgets but research, institutions, expertise, and training the next generation.
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki[Not loaded yet]
- Congrats to everyone involved in the long COVID GWAS - diligence and perseverance paid off with an association near FOXP4 that increases risk (OR=1.63; P=1.76e-10). nature.com/articles/s41...
- Wow, Regeneron bought 23andMe for $256 million. www.cnn.com/2025/05/19/b...
- 🤯Wow, Kári Stefánsson stepped down from deCODE - end of an era. www.ruv.is/english/2025...
- Reposted by Jack Kosmicki[Not loaded yet]
- Carl Sagan was really ahead of his time. Absolutely prescient quote from one of his final interviews.
- It's incredible how far we have come from saturated mutagenesis of 1 gene (BRCA) in 2018 to now 500 protein domains. www.nature.com/articles/s41...
- As the topic of batch covariates from sequencing data came up recently, here was a fun little look at differences in read depth by DNA source (blood or saliva) and sequencing center in All of Us. It often isn't a bad idea to include batch covariates to account for some of these differences!
- Reposted by Jack KosmickiBreast cancer diagnoses rose by over 300% between 1973 and 1992. Was there a breast cancer epidemic?! NO. Mammograms were implemented for screening in 1976. PREVALENCE is a function of being able to correctly diagnose something. This is also about autism.
- Wow, they even scrapped my papers - must have been desperate for data 😅
- My job might be DOOMed to fail, but I certainly got a laugh out of it 😆 I love it when programmer's have a sense of humor (in this case, the people behind Cromshell).
- This is by far the best take on the dire wolves I've heard yet.
- Reposted by Jack Kosmicki[Not loaded yet]
- 23andMe files for bankruptcy 😬 investors.23andme.com/news-release...
- Reposted by Jack KosmickiIn his confirmation hearing now, Bhattacharya was asked by Senator Cassidy whether he thinks there's a need for NIH to study the myth that vaccines cause autism. Bhattacharya says he doesn't believe there's a link, but leaves door open to wasting taxpayer money & scientists' time relitigating this.
- Reposted by Jack KosmickiReally excited to share our next population-scale WGS work preprint. Here, we analyse three anthropometric traits in nearly 700,000 individuals (discovery UKB ~450K, replication AoU). We show, for these traits, that common and rare variant heritability is convergent www.biorxiv.org/content/10.1...
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki“The thing that I can’t get over is that the actual richest man in the world directed my fucking firing. I make $50k a year and work to keep drinking water safe. The richest man in the world decided that was an expense too great for the American taxpayer.”
- 🤦 🤦 🤦 bbc.com/news/article...
- Per the readme, "Variants with a p-value of zero actually had a p-value that exceeded the number of significant digits possible in R (p<1E-321) but can be recalculated using the effect size and standard error provided." Guess it affected more than just the sumstats 😬
- How common is reinfection with SARS-CoV-2 in the omicron world? Basically, no protection from reinfection of omicron SARS-CoV-2 after 1 year 😬 www.nature.com/articles/s41...
- Nice paper showing ~1/540 people have a sex chromosome trisomy from 3 large biobanks (UKB, FinnGen, MVP) comprising 1.5 million individuals along with their phenotypic impact. www.medrxiv.org/content/10.1...
- I've been finding the phrase "the new regime" very useful these days.
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki[Not loaded yet]
- Reposted by Jack Kosmicki[Not loaded yet]
- ☺️So excited for Kristin Tsuo's PhD defense!