Latest from our lab! A proteome-wide association study of cardiovascular diseases in 640,000 participants of multiple ancestries www.medrxiv.org/content/10.6... A truly global collaboration of computational and experimental researchers. h\t @yuxu.bsky.social, Doug Loesch, Henry Taylor among many!

🚨 New preprint from the lab! We’re excited to share “Improving population-scale disease prediction through multi-omics integration” by Ng et al. www.medrxiv.org/content/10.1...

📣📣📣 Excited for our lab's latest preprint, led by Chief Ben-Eghan! www.medrxiv.org/content/10.1... tl;dr We identify protein vQTLs in multiple ancestries then use MVMR to show independent effects of mean & variance on disease, suggesting targeting protein variance could have therapeutic potential.

Wonderful contributions from many... tagging in bsky's I know @cfoguet.bsky.social @xilinjiang.bsky.social @yuxu.bsky.social @sritchie73.bsky.social @iamslambert.bsky.social @aidanbutty.bsky.social @stevesphd.bsky.social

📣📣 Cool job alert! The @bakerresearchau.bsky.social is recruiting up to 2 new PIs in Biomedical Data Science 👉 www.seek.com.au/job/85989791... You’d join an awesome institute in an incredible city (Melbourne, Australia) as well as partnerships like with @cam.ac.uk… you’ll also get to work with me 😁

Super excited to see beginning integration of PredictDB into OmicsPred! We've annotated/deposited the GTExV8 gene expression predictors so they are now available alongside all the other multi-omic predictors at www.omicspred.org More on its way... a wonderful collab with @hakyim.bsky.social & co!

We're working closely with our PredictDB colleagues to bring you an integrated resource for multi-omic predictors We're happy to announce that PrediXcan GTExV8 expression predictors are now annotated and openly available via OmicsPred! www.omicspred.org More info www.omicspred.org/publication/...

Can deep-learning be used for polygenic risk scores? New paper out in @natcomms.nature.com! To find out more read here: www.nature.com/articles/s41... @mikeinouye.bsky.social @aidanbutty.bsky.social @yuxu.bsky.social

Performance of deep-learning-based approaches to improve polygenic scores - Nature Communications

Polygenic scores aim to capture genetic risk but may miss nonlinear genetic and environmental interactions. Here, the authors show that neural networks detect limited nonlinearity and do not outperfor...

nature.com

📣📣 New postdoc in biomedical data science available in our lab!! www.jobs.cam.ac.uk/job/50966/ Flexible funding - many research areas in scope + ideas welcome Check out our group www.inouyelab.org/home And the awesome Cambridge research environment! @vpd-hlri.bsky.social @uniofcam.bsky.social

Looks like a nice resource: * Array genetic data for 80,638 Japanese children * 1,163 child health and developmental traits (e.g. food allergy, anthropometric, developmental) * Parental environmental exposures www.medrxiv.org/content/10.1...

Interesting week reads and resources on human genetics, omics & precision medicine🧵 1⃣From a dog GWAS to a new obesity target! DENND1B variants were the top GWAS signal for obesity in 241 Labrador retrievers—DENND1B variants were also associated with BMI in UK Biobank 🔗 science.org/doi/10.1126/...

Wonderful to see our collaboration @astrazeneca.bsky.social @dphpc.bsky.social out! Identification of plasma proteomic markers underlying polygenic risk of T2D and related comorbidities @naturecomms.bsky.social www.nature.com/articles/s41... Well done Doug Loesch, Dirk Paul, Abhishek Nag and co!

Introducing a major upgrade to OmicsPred platform (www.omicspred.org) — a resource to enhance the accessibility and usability of genetic scores for multi-omic traits and their phenotypic associations. (1/N)

A month ago we @vevotherapeutics.bsky.social announced that we have generated the largest single-cell perturbation atlas in history, Tahoe-100M. Today, we announce that we will fully open-source Tahoe-100M in Feb, as part of a collaboration with NVidia health to train cell state models.

I wrote about how polygenic heritable conditions present in families and the liability threshold model. This has some counterintuitive implications for considering the risk of a condition in offspring as well as the impact of multi-generational selection. A 🧵:

What happens to heritable conditions across generations?

some counterintuitive properties of the polygenic liability threshold model

open.substack.com

📣WOOT! Just out from the lab: Pitfalls of machine learning models for protein-protein interaction networks academic.oup.com/bioinformati... We benchmark various approaches then dig into why they do well and not so well. Led by @loiclnlg.bsky.social (also of Green Algorithms fame)!

For my debut to bluesky, I am pleased to announce a new preprint: medrxiv.org/content/10.1... Here, we show that NMR biomarker scores combined with PRSs and SCORE2 may have moderate population health benefits for 10-year CVD risk prediction prevention Thread 1/3:

📣Cool new print from Scott Ritchie & the lab (www.inouyelab.org)! Cardiovascular risk prediction using metabolomic biomarkers & polygenic risk scores: A cohort study & modelling analyses medrxiv.org/content/10.1... tldr Integrating NMR scores & PRSs with SCORE2 may have moderate pop health benefit