@konradjk.bsky.social
Genomicist, computational biologist. Assistant professor @ MGH, HMS. Associate member @ Broad Institute
@konradjk.bsky.socialThanks to all involved, especially Sophie who has really driven this project, and the participants and funders of All of Us, UK Biobank, and FinnGen. Happy to hear everyone’s thoughts!
@konradjk.bsky.socialWhere accuracy is a bit reduced (e.g. T2D), they’re not explained by ancestry, age, BMI, sex, or smoking status. Instead, phenotype definition matters: hospital data-trained models vs PCP differ, as does requiring two codes.
@konradjk.bsky.socialWe built models of 15 diseases (prevalent/classification, and 5-year incident/prediction) in UK Biobank, and applied them to All of Us and FinnGen. For the most part, accuracy is preserved for most models
@konradjk.bsky.socialThis was a great collaboration with the All of Us Research Program, who generated and released ~10,000 proteomes from ancestrally-diverse individuals: support.researchallofus.org/hc/en-us/art...
@konradjk.bsky.socialNew preprint from the lab from Sophie Parsa, showing how proteomic disease prediction models are accurate and generalizable across environments/continents: www.medrxiv.org/content/10.6...
@masakanai.bsky.socialReposted by @konradjk.bsky.socialOur single-nucleus immune multiome atlas is now out in @nature.com! 🧬
10M PBMCs from 1,108 @finngen.bsky.social donors recruited by Finnish Blood Service, profiled at @broadinstitute.org with chromatin accessibility and gene expression in the same nuclei, to trace how disease variants act 🧵👇
@konradjk.bsky.socialMy favorite part of this (and I think it’s been fixed) is that if you reply to that email, it sends a message on that channel. So when someone leaves an institution, their auto reply replies to every single message… including on the #general channel…