Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →mims-harvard/tooluniverse-gwas-drug-discovery
Transform GWAS signals into drug targets and repurposing opportunities. Connects GWAS-significant loci to causal genes via fine-mapping/eQTL, then to druggable proteins via DGIdb/OpenTargets, then to existing drugs via ChEMBL.
Passed 9 of 16 test cases with the skill, and 7 of 16 without it.
and 3 more in the full report
5 fixed8 passed either way7 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
I am writing a Python script using the ToolUniverse library to pull GWAS Catalog associations for type 2 diabetes. Should I pass `trait='type 2 diabetes'` and `p_value_threshold=5e-8` directly into the query function?
| Checked | Without skill | With skill |
|---|---|---|
| Specifies disease_trait as the correct argument name instead of trait. | Missed | Passed |
| States that p-value filtering must be performed on returned results because no p-value threshold parameter exists in the function call. | Missed | Passed |
The instructions your model reads when the skill loads.
Transform genome-wide association studies (GWAS) into actionable drug targets and repurposing opportunities.
IMPORTANT: Always use English terms in tool calls. Respond in the user's language.
This skill bridges genetic discoveries from GWAS with drug development by:
Key insight: Targets with genetic support have 2x higher probability of clinical approval (Nelson et al., Nature Genetics 2015).
GWAS-to-drug translation succeeds when you think causally. A genetic association provides causal direction that observational data cannot: if a loss-of-function variant protects against disease, an inhibitor of that gene's product is the hypothesis to test. The direction of effect (LOF vs. GOF) determines whether you need an inhibitor or an agonist — get this wrong and the drug works backwards. GWAS effect sizes are small (odds ratios of 1.1–1.5 are typical), but the drug effect may be much larger or smaller than the genetic effect; the genetic signal validates the target, not the dose. Always integrate multiple lines of evidence (eQTL colocalization, pQTL, L2G score) before committing to a target, because many GWAS variants tag the causal gene only indirectly.
LOOK UP DON'T GUESS: Do not assume which gene a GWAS variant implicates — use OpenTargets_get_variant_credible_sets or gwas_get_associations_for_snp to get the actual mapped gene and L2G score. Do not guess the direction of effect, odds ratio, or whether a drug already exists for the target; always query the tools.
Input: Disease/trait name (e.g., "type 2 diabetes", "Alzheimer disease")
Process: Query GWAS Catalog for associations, filter by significance (p < 5x10^-8), map variants to genes, aggregate evidence.
Tools:
gwas_get_associations_for_trait - Get associations by diseasegwas_search_associations - Flexible searchgwas_get_associations_for_snp - SNP-specific associationstest cases passed, vs 7 / 16 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.