Finding the Right Target: How BioMate Turns Human Genetics Into Drug Discovery
Newsroom · BioMate Showcase

Finding the Right Target: How BioMate Turns Human Genetics Into Drug Discovery

Drug DevelopmentAug 17, 20268 min readBioMate Team

The single best-quantified predictor of clinical success is human genetic support — drug mechanisms backed by genetics reach approval at roughly 2–2.6× the base rate (Nelson 2015; Minikel 2024). Our Target Identification survey maps the whole field. This article is the shorter question: what does BioMate actually do with it?

The landscape, in one paragraph

Modern target discovery is a pipeline: a GWAS points at a genomic locus, not a gene; locus-to-gene ML, colocalization with eQTL/pQTL data, and cis-Mendelian randomization narrow it to a causal gene and its direction of effect; CRISPR screens and DepMap turn that into a tested dependency; and platforms like Open Targets score everything together. The winning programs chain genetics, single-cell, perturbation, and structure rather than betting on one model.

What BioMate does

BioMate is the execution engine for exactly that genetics-first recipe — run through plain-English requests on managed AWS Batch, with quality-graded, audit-ready results.

  • Genetics-anchored target pipelines. A five-stage chain — GWAS gene mapping → eQTL colocalization → single-cell GWAS → Mendelian randomization → Open Targets integration — runs end-to-end. On a rheumatoid-arthritis run over 932,549 patients it prioritized TYK2, JAK1, IL6R, PTPN22, and CD20 in 15–30 minutes.
  • Functional-genomics readouts. Single-cell and expression workflows surface disease-associated cell states and effector genes, with cell-type specificity that doubles as an on-target safety filter.
  • From target to molecule. ~215 tested drug-discovery workflows span docking/QSAR/RBFE, a Pareto multi-parameter optimizer, and an ADMET auto-loop that reruns with corrected parameters when a QC gate fails (hERG, DILI).
  • Direction & developability. Genetic direction-of-effect (inhibit vs. activate) and ADMET/PBPK developability carry through to an IND-style evidence package.
Plain-English request→GWAS → gene mapping→eQTL colocalization→Mendelian randomization→Open Targets scoring→ADMET / developability→Ranked targets + report
Figure. How BioMate runs genetics-anchored target discovery — one plain-English request drives the full chain on AWS Batch.
The bottom line. The genetics-first pipeline the evidence says de-risks programs usually sits behind a dedicated bioinformatics team. BioMate puts it one plain-English request away — from a GWAS to ranked, developability-checked targets with a signed audit trail — so a wet-lab biologist runs in an afternoon what used to take weeks.

References

  1. Nelson MR, et al. The support of human genetic evidence for approved drug indications. Nat Genet 2015;47:856. doi:10.1038/ng.3314
  2. Minikel EV, et al. Refining the impact of genetic evidence on clinical success. Nature 2024;629:624. doi:10.1038/s41586-024-07316-0
  3. Mountjoy E, et al. An open approach to prioritize causal variants and genes at GWAS loci (locus-to-gene). Nat Genet 2021;53:1527. doi:10.1038/s41588-021-00945-5

Go deeper: the full survey

This showcase is one half of a pair. For the complete, citation-backed map of the field — every method, its strengths and limits, and what is actually winning — read the Drug Target Identification survey.

Try BioMate free
Read the survey: Drug Target Identification → · All surveys: Research Surveys → · More: Newsroom