🚨NEW PREPRINT! Interested in Mendelian randomization and proteomics? 🧬 We created a comprehensive multi-ancestry proteome-phenome-wide atlas of protein-disease associations in European, African, and East Asian ancestries! Preprint: www.medrxiv.org/content/10.1... Website: broad.io/protein_mr_a... - ThreadSky

chenyangsu.bsky.social • 100 days ago

🚨NEW PREPRINT!
Interested in Mendelian randomization and proteomics? 🧬

We created a comprehensive multi-ancestry proteome-phenome-wide atlas of protein-disease associations in European, African, and East Asian ancestries!

Preprint: https://www.medrxiv.org/content/10.1101/2024.10.17.24315553v2

Website: https://broad.io/protein_mr_atlas

Comments

chenyangsu.bsky.social•100 days ago

As exposures, we used a total of 7 proteomics cohorts including a novel East Asian ancestry cohort from Kyoto university.

As outcomes, we manually curated the largest GWAS for European and African ancestries, while outcomes from Biobank Japan were used for East Asian ancestry.

chenyangsu.bsky.social•100 days ago

By including proteomics cohorts from non-European ancestries, we were able to test an additional 119 proteins that were unique in African ancestry and 17 in East Asian ancestry due to allele frequency differences.

chenyangsu.bsky.social•100 days ago

We performed extensive sensitivity analyses and colocalization analyses with PWCoCo and SharePro and identified 3,949, 56, and 325 putatively causal protein-phenotype pairs in European, African, and East Asian ancestries, respectively.

chenyangsu.bsky.social•100 days ago

We built an interactive web browser of all our results which are hosted on the Common Metabolic Diseases Knowledge Portal!

chenyangsu.bsky.social•100 days ago

Naturally, we would want to know which of these proteins implicated in diseases are druggable. To answer this, we incorporated three drug databases.
💊The druggable genome
💊DrugBank
💊 https://opentargets.bsky.social.

chenyangsu.bsky.social•100 days ago

64.5% of proteins implicated in at least one phenotype overlapped with at least one database.
Specifically,
💊 55.8% of proteins overlapped with the druggable genome
💊 33.8% had approved or investigational drugs in DrugBank
💊18.4% overlapped with the Open Targets Platform.

Comments

Posting Rules

Reply