Agentic Genomics: Leverage the Power of AI in Genomics
Agentic Genomics: Hands-on with AI for Variant Interpretation and GWAS
The variant interpretation workshop that went viral on LinkedIn is now a live session, with a brand new GWAS module. Join online via Microsoft Teams.
Two hands-on sessions, both running entirely in Google Colab. No installation, no terminal, no API keys, no cost. Just a Gmail account and a browser.
SLIDES
Introduction: https://clawbio.ai/workshop-slides.html Variant Interpretation: https://clawbio.ai/workshop-slides.html GWAS: https://clawbio.ai/workshop-gwas-slides.html
SESSION 1: Variant Interpretation (30 min)
Annotate a real human genome (Dr Manuel Corpas's, published under CC0 in 2013) and discover clinically relevant findings: Factor V Leiden carrier status, CFTR cystic fibrosis carrier, warfarin dose sensitivity, APOE Alzheimer's risk, and hereditary hemochromatosis.
What you'll do:
Run a complete variant interpretation pipeline in Google Colab
Annotate variants using Ensembl VEP, ClinVar, and gnomAD
Generate pharmacogenomics recommendations using CPIC guidelines
Classify variants using ACMG criteria
Materials: https://docs.clawbio.ai/tutorials/variant-interpretation-workshop/
SESSION 2: GWAS with ClawBio (30 min)
Run a genome-wide association study using publicly available summary statistics. Query variants across nine federated databases, compute polygenic risk scores, and perform fine-mapping to identify causal variants.
What you'll do:
Look up GWAS variants across GWAS Catalog, Open Targets, gnomAD, GTEx, PheWAS, and more
Compute polygenic risk scores from published PGS Catalog scores
Fine-map a GWAS locus using SuSiE to identify credible sets
Explore cross-ancestry differences in variant frequencies and effect sizes
Materials: https://docs.clawbio.ai/tutorials/gwas-workshop/
This session builds on the research agenda developed with Prof Segun Fatumo (QMUL/PHURI) around democratising genomics for the Global South: removing the barriers of HPC infrastructure, data access, and training so that researchers anywhere can run publication-quality analyses.
Q&A and Discussion (30 min)
Who should attend
Researchers, bioinformaticians, clinicians, and students interested in genomics and AI. No prior AI or programming experience required. Bring a laptop with a Google account.
About ClawBio
ClawBio is an open-source skill library for reproducible agentic bioinformatics. 488+ GitHub stars, 39 bioinformatics skills, 13 contributors.
GitHub: https://github.com/ClawBio/ClawBio
Whatsapp: https://chat.whatsapp.com/FInuUxJa00G7Ql3T28kW4I?mode=gi_t
Discord (with RoboTerri): https://discord.gg/DHpmptyv