SCI Skills

Data & ReproDrug Discovery & CheminformaticsFreedomIntelligence/OpenClaw-Medical-SkillsData & Reproduction

PO

popeve-variant-predictor-agent

Maintainer FreedomIntelligence · Last updated April 1, 2026

Variant pathogenicity prediction using EVE population-based evolutionary models.

OpenClawNanoClawAnalysisReproductionpopeve-variant-predictor-agent🧠 bioos extended suitedrug discovery & designvariant

Original source

FreedomIntelligence/OpenClaw-Medical-Skills

https://github.com/FreedomIntelligence/OpenClaw-Medical-Skills/tree/main/skills/popeve-variant-predictor-agent

Maintainer: FreedomIntelligence
License: MIT
Last updated: April 1, 2026

Skill Snapshot

Key Details From SKILL.md

2 min

Key Notes

The PopEVE Variant Predictor Agent leverages the PopEVE deep learning model from Harvard Medical School to predict pathogenicity of genetic variants. PopEVE analyzes evolutionary conservation, protein structure, and population frequency to identify disease-causing variants, having identified over 100 previously unrecognized variants responsible for undiagnosed rare genetic diseases.
When predicting pathogenicity of missense variants genome-wide.
For rare disease diagnosis with variants of uncertain significance (VUS).
To prioritize candidate variants in exome/genome sequencing.
When interpreting novel variants not in ClinVar or literature.

Source Doc

Excerpt From SKILL.md

Core Capabilities

Pathogenicity Prediction: Score any missense variant for disease likelihood.
VUS Resolution: Reclassify variants of uncertain significance.
Rare Disease Diagnosis: Identify causal variants in undiagnosed patients.
Population-Aware Scoring: Account for ancestry-specific variant frequencies.
Protein Context Analysis: Integrate structural and functional domains.
Batch Variant Scoring: Process thousands of variants efficiently.

Model Architecture

Component	Description	Data Source
Evolutionary Module	Deep sequence alignment	UniRef90, 250M seqs
Structural Module	AlphaFold2 structures	200M+ structures
Population Module	gnomAD frequencies	800K+ individuals
Clinical Module	ClinVar training	100K+ classifications
Integration	Multi-task neural network	Combined features

Scoring Thresholds

PopEVE Score	Interpretation	Suggested Action
> 0.9	Likely Pathogenic	High priority
0.7 - 0.9	Possibly Pathogenic	Review carefully
0.3 - 0.7	Uncertain	Additional evidence needed
0.1 - 0.3	Possibly Benign	Lower priority
< 0.1	Likely Benign	Deprioritize

Use cases

When predicting pathogenicity of missense variants genome-wide.
For rare disease diagnosis with variants of uncertain significance (VUS).
To prioriti.

Not for

Do not rely on this catalog entry alone for installation or maintenance details.

Upstream Related Skills

AlphaMissense_Agent - For AlphaMissense predictions
DiagAI_Agent - For clinical diagnosis support
ACMG_Classifier_Agent - For ACMG classification
Pharmacogenomics_Agent - For drug-gene variants

Related skills

Related skills

Back to directory

AA

Data & ReproDrug Discovery & Cheminformatics

aav-vector-design-agent

AAV vector design: capsid selection, promoter optimization, payload capacity.

OpenClawNanoClawAnalysis

FreedomIntelligence/OpenClaw-Medical-SkillsView

AG

Data & ReproDrug Discovery & Cheminformatics

agentd-drug-discovery

AgentD autonomous drug discovery: target identification, hit finding, ADMET optimization.

OpenClawNanoClawAnalysis

FreedomIntelligence/OpenClaw-Medical-SkillsView

BI

Data & ReproDrug Discovery & Cheminformatics

bio-clinical-databases-hla-typing

Call HLA alleles from NGS data using OptiType, HLA-HD, or arcasHLA for immunogenomics applications. Use when determining HLA genotype for tr…

OpenClawNanoClawAnalysis

FreedomIntelligence/OpenClaw-Medical-SkillsView

BI

Data & ReproDrug Discovery & Cheminformatics

bio-clinical-databases-pharmacogenomics

Query PharmGKB and CPIC for drug-gene interactions, pharmacogenomic annotations, and dosing guidelines. Use when predicting drug response fr…

OpenClawNanoClawAnalysis

FreedomIntelligence/OpenClaw-Medical-SkillsView