Ancestry aware mr
Skill Alim430/bioresearch-agent/skills/biomedical/ancestry-aware-mr
Executable biomedical workflows for AI assistants — literature analysis, biomarker discovery, and Mendelian randomization through reproducible agent skills.
npx -y skills add Alim430/bioresearch-agent --skill ancestry-aware-mrAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Run cross-ancestry Mendelian randomization with correlated-pleiotropy modeling (CAUSE-like) and mixture-model pleiotropy detection (MRMix-like), plus per-ancestry IVW, cross-ancestry meta-analysis (FE/RE), and portability assessment. Use when the user needs to test whether a causal effect estimated in one ancestry transfers to others, or to distinguish true causation from correlated horizontal pleiotropy. Mock mode validates the MR + pleiotropy pipeline; live mode would use BBJ/FinnGen/TPMI/UKB summary statistics.
SKILL.md
3.5 KB, as published. Nobody here has run it
BioResearch Agent — Ancestry-Aware MR Skill
Capability
Runs cross-ancestry Mendelian randomization with pleiotropy-aware methods:
- Per-ancestry IVW — inverse-variance-weighted MR per ancestry with Cochran's Q heterogeneity test, producing ancestry-specific causal effect estimates and confidence intervals.
- Cross-ancestry meta-analysis — fixed-effects (FE) and random-effects (RE, DerSimonian-Laird) meta-analysis across ancestries, with Cochran's Q and I² statistics to quantify cross-ancestry heterogeneity.
- CAUSE-like model — (Morrison et al. 2020) EM algorithm modeling both correlated and uncorrelated horizontal pleiotropy. Tests H0 (no causal effect, pleiotropy free to vary) vs H1 (causal effect + pleiotropy) via likelihood-ratio test. Distinguishes true causation from correlated pleiotropy — the key confound in standard MR.
- MRMix-like model — (Wang et al. 2020) three-component mixture model (causal / pleiotropic / null) via EM, estimating the proportion of pleiotropic instruments and adjusting the causal estimate accordingly.
- Portability assessment — evaluates whether the causal effect estimated in a reference ancestry (e.g., EUR) transfers to others: direction consistency, significance consistency, heterogeneity (I²), EUR-centric bias, and a composite transferability score (0–1).
Returns per-ancestry MR + meta-analysis + CAUSE/MRMix + portability report, not a definitive causal claim.
Run
bioresearch run ancestry-aware-mr --n-snps 200 --n-instruments 40 --true-effect 0.30 --seed 42 --output-dir outputs/ancestry-mr
Outputs (in --output-dir)
amr_per_ancestry_results.csv— per-ancestry IVW results (ancestry, beta, se, CI, p, Q, p_het)amr_cross_ancestry_meta.csv— FE + RE meta-analysis (method, beta, se, CI, p, Q, I², tau²)amr_cause_results.csv— CAUSE-like model per ancestry (theta, eta, sigma, LRT_stat, p_value)amr_mrmix_results.csv— MRMix-like model per ancestry (theta, pi_causal, pi_pleiotropic, pi_null)amr_portability.csv— portability assessment (direction_consistency, I², transferability_score)amr_forest_plot.png— forest plot of per-ancestry + meta-analysis causal estimatesamr_evidence_package.json— reproducible Evidence Package (provenance + parameters + grade)
Note
This skill dispatches to the framework's ancestry-aware-mr workflow / demo_ancestry_aware_mr.py.
It adds no analysis of its own; all computations run in the workflow modules. By default uses
simulated multi-ancestry GWAS with known causal effect and controlled pleiotropy rates to
validate the MR + pleiotropy detection pipeline — real-data deployment would use BBJ (EAS),
FinnGen (EUR), TPMI (SAS), and All of Us (AMR/AFR) summary statistics with ancestry-matched LD panels.
Evidence grade is C (methodology validation). Part of Phase 3a (cross-ancestry MR, CPU-only).