Claude Skills by ma-compbio-lab

Use this skill to search Crossref metadata when you need DOI-oriented citation records from the official REST API. Do not use it when you need full text or publisher-specific access controls.

Extract likely dataset, code, and package links from locally prepared paper text or markdown using deterministic URL and keyword rules.

Use this skill to search life-science literature through Europe PMC when you need publication metadata oriented toward biomedical method triage. Do not use it when you need offline execution or cluster-scale batch processing.

Extract figure and table captions from locally prepared paper text using deterministic caption patterns.

Use this skill to query OpenAlex, Europe PMC, Crossref, and PubMed for the same topic and build a normalized literature brief. Do not use it as a full systematic-review workflow or citation-analysis engine.

Use this skill to search PubMed through NCBI E-utilities when you need official Entrez-backed literature discovery and summary metadata. Do not use it when you need full text or network-free execution.

Use this skill to look up a seed paper in OpenAlex and summarize both upstream references and downstream citing works.

Use this skill to run fast literature triage against OpenAlex when you need public paper metadata, citation-linked context, and structured JSON output. Do not use it when you need full text, proprietary databases, or offline execution.

Use this starter when a task lands in the `Protocol and workflow extraction` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Reproducibility cue extraction` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Scientific summarization` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Rank a small set of candidate papers for triage using deterministic citation, recency, and query-overlap heuristics inspired by Semantic Scholar recommendation workflows.

Mine likely review, survey, and meta-analysis papers from a local metadata set using deterministic title and abstract heuristics aligned with Semantic Scholar review-search workflows.

Use this starter when a task lands in the `Active learning` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to generate a deterministic toy posterior and summarize uncertainty diagnostics with `ArviZ`.

Use this starter when a task lands in the `Clustering` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to estimate a deterministic toy average treatment effect with DoWhy and summarize a simple placebo refutation.

Use this starter when a task lands in the `Foundation models for biology / chemistry / science` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Multimodal fusion` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to run a deterministic Optuna optimization loop on a small synthetic objective and inspect the best trial summary.

Use this skill to generate a tiny full-factorial design from bounded factor definitions and summarize the resulting experiment table.

Use this skill to fit a tiny Bayesian linear regression with PyMC and summarize posterior means plus credible intervals.

Use this starter when a task lands in the `Representation learning` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to run a small deterministic hypothesis-testing workflow over two toy groups and summarize the results.

Use this starter when a task lands in the `Simulation-based inference` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to embed a tiny labeled feature matrix with UMAP and summarize the resulting 2D geometry.

Use this starter when a task lands in the `Binding site analysis` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Fold comparison` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Protein complex metadata` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to fetch a canonical RCSB PDB entry record by accession when you need quick structural metadata. Do not use it for search ranking or full structure parsing pipelines.

Use this skill to search the Protein Data Bank through the official RCSB Search API when you need structure identifiers from a free-text query. Do not use it when you need structure download or coordinate parsing.

Use this starter when a task lands in the `Sequence/structure alignment` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Structure parsing and cleanup` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Structure-to-sequence cross-linking` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Visualization` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this starter when a task lands in the `Causal network analysis` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to run deterministic local custom enrichment with Bioconductor clusterProfiler and user-supplied TERM2GENE tables. Do not use it for remote annotation downloads or organism database lookups.

Use this skill to run deterministic toy or local custom preranked gene-set enrichment with Bioconductor fgsea. Do not use it for remote annotation lookups or large cohort-scale pathway workflows.

Use this skill to build a deterministic toy biological interaction graph and summarize basic network properties with NetworkX.

Use this skill to propagate signal from one or more seed nodes over a small weighted interaction graph with personalized PageRank.

Use this skill to search public Gene Ontology terms through the official QuickGO API and return compact ontology summaries that are easy to inspect or feed into downstream enrichment workflows.

Use this skill to fetch a concise summary for a known Reactome event or pathway stable ID from the official Reactome Content Service. Do not use it for full pathway traversal, graph expansion, or unsupported Reactome endpoints.

Use this skill to submit a short identifier list to Reactome Analysis Service and return compact pathway enrichment summaries. Do not use it for full downstream statistical analysis or unsupported identifier normalization.

Use this skill to submit a short identifier list to the official Reactome Analysis Service and return a compact ranked pathway table.

Use this skill to traverse the official Reactome pathway hierarchy and summarize the ancestor chain, direct children, and descendant count for a target stable ID.

Use this starter when a task lands in the `Regulatory network inference` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to query the official STRING API for high-confidence interaction partners of one gene or protein identifier and summarize the returned evidence in compact JSON.

Use this starter when a task lands in the `Bulk RNA-seq QC and normalization` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.

Use this skill to search the public CELLxGENE Census dataset catalog by keyword and summarize the largest matching atlas datasets.

Use this starter when a task lands in the `Isoform / transcript-level analysis` frontier leaf and the repository has curated resources but no dedicated runtime implementation yet.