Brandon Coombes

A fantastic study by @xiangruimeng158 @Kkuchenbaecker and colleagues was recently published in Nature Genetics. The paper studied genetic risk for major depression across 1.8 million individuals with 54% from non-European ancestry. link: nature.com/articles/s4158… 🧵below

In the latest edition of "Huh, I didn't expect that to work", our latest paper shows that LLMs can outperform existing methods for identifying causal genes in genome-wide association studies. 🧵 medrxiv.org/content/10.110…

It pains me to see facile critiques of GWAS on here from our clinical/biostats friends while the many actually good reasons to be critical of GWAS get little attention. So here's a thread on what GWAS does, what critics get wrong, and where GWAS is genuinely still lacking. 🧵:

Mendelian Randomisation (MR) is going to hell in a handcart, with an exponential weekly increase of papers with Mendelian Randomisation in the title, most of them nonsense. 25th April 4.30pm online or in person @ucl I'll talk about what should be done forms.office.com/pages/response…

Lots of discussion of Aw et al. [biorxiv.org/content/10.110…] in our journal club this week. Potentially important implications for polygenic score (PGS) analysis and disease architecture. Curious what other people think. Some thoughts below:

The publication of the whole genomes from the US @AllofUsResearch cohort is great to see, but the choice of how to represent an overview of the genetic relationships has (rightly) drawn controversy, in particular how the concepts of ethnicity and race are mapped to it.

A few thoughts on the recent set of papers torture testing genomic deep learning for predicting individual-level gene expression [ pubmed.ncbi.nlm.nih.gov/38036790/ , pubmed.ncbi.nlm.nih.gov/38036778/ , pubmed.ncbi.nlm.nih.gov/38036789/ ]. First a brief summary 🧵: twitter.com/SashaGusevPost…

Our new study of cannabis use disorder (N>1,000,000) led by @dflevey1 is now online at @NatureGenet. Article here: go.nature.com/49Nj1Ou Sumstats will be made available here: - dbGAP (bit.ly/47kgf1u) - Gelernter Lab Webpage (bit.ly/3ETJWJU)

Why do family history and polygenic scores explain independent variance for complex traits? Is it because family history is an 'environmental instrument'? Or are these just two very noisy genetic predictors? See our new paper for a deep dive. medrxiv.org/content/10.110…

New in @ScienceMagazine: PrimateAI-3D, a ML classifiers trained on protein folding + primate genetic data to identify damaging DNA variants in humans, created ‘rare variant polygenic scores’ that explain outlier values for cholesterol and other traits science.org/doi/10.1126/sc…

GWAS folks: why aren't we surprised to see manhattan plots end at chr 22 while we usually have 23 pairs of chromosomes? In this review @AJHGNews, we show that there is a 7x difference in... authors.elsevier.com/sd/article/S00…

Today's theme is #COMORBIDITY. Some recent studies have used genetics to unpack the comorbidity between #depression and other psychiatric conditions. Read more below -

Genetic scores for predicting levels of biomolecules validated across diverse ancestries enable insights into disease biology. omicspred.org enables researchers to predict molecular traits from genetics @minouye271 Yu Xu nature.com/articles/s4158… nature.com/articles/d4158…

“Obesity is genetic” If you were just triggered, then this thread is for you. If “but the human genome can’t evolve in a generation!” has passed your lips in the past few months, then this thread is for you. 🧬This is why obesity is genetic. And the environment is to blame.

Just Online: rdcu.be/c8efR. We conducted the largest GWAS of addiction to date by combining information across substance use disorders. A thread on a GWAS of addiction in >1 million individuals across two ethnicities #GWAS #MedTwitter #PsychTwitter #Addiction (1/9)

Now online at Nature Mental Health (@NatMentHealth): Multivariate genome-wide association meta-analysis of over 1 million subjects identifies loci underlying multiple substance use disorders led by @AlexanderHatoum of @PGCgenetics.

While I appreciate the sentiment, I am not a fan of such blind rules. Let me start with a 🧵 of some major human genetic discoveries (many translated to therapeutics) made with very few samples. twitter.com/ewanbirney/sta…

All source code available in github, here: github.com/FINNGEN/lavaa-… Give it a try anywhere you might use a pheWAS. I'd love to see other examples where this sort of view may aid in interpretation of a locus.

Applications of PRS in the future will require better understanding gene-environment interplays. New preprint from our lab on joint modeling of PRS-E correlations and interactions in case-control studies. biorxiv.org/content/10.110…