メニュー
Functional analysis of a gene list — batch summaries, pathway mapping, protein interactions, tissue expression, and phenotype associations
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/syntheticgio/AIBioSkills --skill gene-list-agentコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SOC 職業分類に基づく
SKILL.md を表示中
| name | gene-list-agent |
| description | Functional analysis of a gene list — batch summaries, pathway mapping, protein interactions, tissue expression, and phenotype associations |
Perform a functional analysis of the gene list: $ARGUMENTS
Use the MCP tools to characterize this gene set — summarize each gene, find shared pathways, map interactions, assess tissue expression patterns, and identify phenotype associations. Follow the steps below. If a step fails for a specific gene, note the gap and continue.
Extract gene symbols from the input. The user may provide:
TP53, BRCA1, EGFR, PTENTP53 BRCA1 EGFR PTENNormalize all symbols to uppercase. If more than 20 genes are provided, work with the first 20 and note the truncation.
batch_gene_summary with all gene symbols (comma-separated) and taxon human.datasets_summary_gene calls for each gene.For each gene (up to 10):
kegg_get_pathway with the gene symbol to find associated pathways.string_get_interactions with all gene symbols joined by commas (species: 9606, required_score: 700, limit: 10).For each gene (up to 10):
gtex_get_expression with the gene symbol.For each gene (up to 5 — focus on the most interesting or least familiar):
uniprot_search with the gene symbol and organism_id:9606 (reviewed: true).uniprot_get_protein with the accession to get GO terms and function descriptions.clinvar_search with each gene symbol to check for pathogenic variants.For each gene (up to 10):
hpo_search with the gene symbol (category: search) to find associated phenotypes.# Gene List Functional Analysis
## Input
- Number of genes analyzed: N
- Gene symbols: [list]
## Gene Summary Table
| Gene | Full Name | Chr | Type | ClinVar Pathogenic Variants |
|------|-----------|-----|------|-----------------------------|
| TP53 | Tumor protein p53 | 17 | protein-coding | 1,247 |
| ... | ... | ... | ... | ... |
## Shared Pathway Analysis
### Top Shared Pathways
Pathways containing 2+ genes from the input list:
| Pathway | KEGG ID | Genes in List | Total Genes in Pathway |
|---------|---------|---------------|----------------------|
| p53 signaling | hsa04115 | TP53, BRCA1, PTEN | — |
| ... | ... | ... | ... |
### Per-Gene Pathway Summary
Brief note on unique pathways for each gene not shared with others.
## Protein Interaction Network
- Direct interactions between input genes (from STRING)
- Key hub genes (genes with most connections)
- Notable interaction partners outside the input list
## Tissue Expression Patterns
- Shared high-expression tissues across the gene set
- Tissue expression summary table:
| Tissue | Genes Highly Expressed | Notable |
|--------|----------------------|---------|
## Functional Themes
- Recurring GO Biological Process terms (enriched themes)
- Recurring GO Molecular Function terms
- Common subcellular localizations
## Phenotype Associations
- Recurring HPO phenotype terms across the gene set
- Disease associations shared by multiple genes
## Clinical Summary
- Genes with highest ClinVar pathogenic variant burden
- Shared disease associations
## Interpretation
Brief synthesis:
- What biological theme(s) unite this gene set?
- Are they co-expressed in specific tissues?
- Do they share pathway memberships or interact directly?
- What clinical significance does this gene set have collectively?
## Data Sources
List which databases were queried and their success/failure status.
Keep the report factual — only include data returned by the tools. For pathway and enrichment analysis, clearly state that you are reporting observed overlaps, not formal statistical enrichment (no p-values). If a gene returns no data from a source, note it in the relevant section.
Full clinical variant workup — gnomAD population frequency, ClinVar significance, protein domain impact, AlphaFold structure context, and PubMed literature
Compare a gene across multiple species — find orthologs, retrieve sequences, compute alignments, and summarize conservation
Druggability assessment — protein function, 3D structures, ligand-bound PDBs, binding sites, interaction network, disease associations, and literature