bioinfor-claw

Pan-cancer analysis for a gene across all 33 TCGA cancer types plus DepMap and CPTAC. Use this skill when the user asks for pan-cancer analysis, cross-cancer comparison, multi-cancer gene report, or wants to analyze a gene across ALL cancer types at once. Generates a comprehensive PDF report with expression, survival (KM curves + log-rank), mutation, copy-number, DepMap cell-line data, and CPTAC protein data. Covers TCGA expression + survival, TCGA mutation/CNA survival, DepMap 26Q1 expression/copy-number/mutations, and CPTAC proteomics.

DepMap analysis for a gene across CANCER CELL LINES (NOT patient samples). Modules: expression, mutation, copy number, essentiality. Data is streamed directly from the DepMap API — no full dataset download needed. Also includes standalone co-expression (depmap_coexpression.py) and co-essentiality (depmap_coessentiality.py) scripts for cell-line correlations. For patient/tumor co-expression use coexpression-for-gene (TCGA/GTEx) instead.

Co-expression in PATIENT SAMPLES (TCGA tumors / GTEx normal tissues). NOT for cell lines — use depmap_coexpression.py for DepMap cell-line co-expression. Fetches real TCGA expression from cBioPortal API (PanCancer Atlas). GTEx requires user-provided expression file.

Run Cox proportional hazards regression on clinical and molecular data. Fits univariate and multivariate Cox models, computes hazard ratios with confidence intervals, tests the proportional hazards assumption, and generates forest plots and Schoenfeld residual plots.

Correlate gene expression or dependency scores with drug sensitivity data from PRISM (DepMap). Identifies drugs whose sensitivity correlates with the gene's expression or essentiality, and generates correlation plots.

Analyze TCGA somatic mutations in a gene via cBioPortal API (PanCancer Atlas). Supports local TCGA MAF file as fallback. Computes mutation frequencies, identifies hotspot residues, classifies mutation types, and generates lollipop plots and summaries.

Retrieve and summarize normal tissue expression for a single human gene using GTEx, including full tissue-level expression, publication-quality barplot output, and a simple expression-pattern classification such as universally expressed, tissue-specific, mixed, or non-expressed.

Query and visualise TCGA expression data for a single gene across cancer cohorts. Supports pan-cancer bar plots, single-cohort boxplots, and tumor-vs-normal comparisons. All data is fetched live from the GDC API — no local files required.

Perform Kaplan-Meier survival analysis for a single gene in a TCGA cohort. Stratifies patients by expression level (high vs low) and tests Overall Survival (OS) and/or Disease-Free Survival (DFS) with log-rank statistics. All data is fetched from the GDC API — no local data files required.

Design a pooled CRISPR sgRNA library for a list of target genes. Designs sgRNAs for each gene, selects top guides based on predicted efficiency and off-target scores, includes non-targeting controls, and outputs an oligo order file ready for library synthesis.

Discover leading PI-led research groups in a scientific field using OpenAlex publication evidence, with ranking based mainly on overall scholarly impact, PI-like authorship patterns, and recent 10-year and 5-year activity summaries.

Generate publication-quality Kaplan-Meier survival curves from time-to-event data. Supports multiple groups, log-rank test, Wilcoxon test, median survival annotation, confidence intervals, at-risk tables, and flexible stratification by continuous or categorical variables.

Route to the correct publication-quality plot sub-skill for volcano plots, heatmaps, box/violin plots, scatter plots, bar charts, MA plots, correlation matrices, and bubble charts from bioinformatics data.

Analyze CRISPR pooled screens using MAGeCK. Accepts a raw count table or FASTQ files. Runs MAGeCK RRA (pairwise test) or MAGeCK MLE (multi-condition). Produces gene-level hit rankings, volcano plots, sgRNA rank plots, and hit summary figures for both positive and negative selection. Supports genome-wide KO screens, CRISPRi, and CRISPRa experiments.

Quality control analysis for pooled CRISPR screens. Accepts a sgRNA read count matrix (guides × samples) and computes library representation, read depth, guide-level QC metrics, replicate correlations, and Gini index. Generates a QC report with plots.

Design base editor sgRNAs for a given gene. Fetches exonic sequence from Ensembl, scans for all NGG (or NG/NGA) PAM sites, filters for guides where a cytosine (CBE) or adenine (ABE) falls in the editing window, scores guides for efficiency and bystander risk, and returns a ranked table. Supports CBE (C→T), ABE (A→G), and dual-base editors for human, mouse, and monkey.

Design pegRNAs (prime editing guide RNAs) and nicking sgRNAs for prime editing. Given a target gene and desired edit (SNV, small insertion, or deletion), identifies PE-compatible protospacers, designs RT templates and PBS sequences, and scores guides by predicted efficiency.

Compare 2 to 6 gene lists for overlap and enrichment. Computes pairwise Jaccard similarity, overlap coefficients, and Fisher's exact test for enrichment relative to a background. Generates Venn diagrams (up to 4 sets), UpSet plots, and an overlap matrix heatmap.

Use when the user wants to run enrichment analysis from a gene list and generate publication-quality bubble plots for Gene Ontology, KEGG, and Reactome terms.

Build and analyze a protein-protein interaction (PPI) network for a gene list using the STRING database API. Returns network edges, computes key network metrics, identifies hub genes, detects modules, and visualizes the network.

Find transcription factors (TFs) that significantly regulate a given gene list. Uses enrichment analysis against curated TF-target databases (ENCODE ChEA3 API, DoRothEA gene sets built-in) and returns ranked TFs with enrichment statistics and network visualization.

Downstream analysis of ATAC-seq or ChIP-seq peak files. Annotates peaks to genomic features, computes signal summaries, finds differential peaks between conditions, and generates QC plots.

Analyze DNA methylation data from Illumina 450K/EPIC arrays or bisulfite sequencing. Computes differential methylation between groups, identifies differentially methylated regions (DMRs), annotates CpGs, and visualizes results.

Analyze mass spectrometry proteomics data (TMT, LFQ, or DIA-NN output). Normalizes protein intensities, filters low-quality measurements, identifies differentially expressed proteins, and generates publication-quality visualizations.

Perform RNA-seq differential expression analysis between two sample groups from a count matrix. Supports DESeq2-style (via pydeseq2), Welch's t-test, and Mann-Whitney U methods. Outputs full DE results, significant gene table, volcano plot, MA plot, and top-gene heatmap. Outputs feed directly into go-analysis-for-gene-list and gsea-for-ranked-gene-list.

Basic single-cell RNA-seq analysis pipeline. Accepts a raw count matrix (cells × genes or genes × cells), performs QC filtering, normalization, highly variable gene selection, dimensionality reduction (PCA + UMAP), Leiden-like clustering, and marker gene identification. Generates UMAP plots and a marker gene heatmap.

Search and track recent preprints from bioRxiv and medRxiv on a given topic. Retrieves preprint metadata, abstracts, and author information via the bioRxiv/medRxiv API. Identifies trending topics, prolific authors, and generates a digest report.

Search PubMed for scientific papers matching a query. Retrieves ranked results with abstracts, journals, citation counts (via Europe PMC), and author information. Filters by date range, journal type, publication type, and study type. Saves a structured TSV and a formatted report.

Curate literature-backed gene or protein lists for a requested function, pathway, molecular class, or regulator type, then normalize them into a structured table or workbook. Use when the user asks for lists such as transcription factors, pathway genes, receptor families, chromatin regulators, methylation readers, or other function-specific gene sets with identifiers and references.

Use when the user wants to retrieve CRISPR sgRNA candidates for one gene or a gene list from the GuidePro genome-wide reference table for human, monkey, or mouse, and return the top guides per gene.

Retrieve and summarize normal tissue expression for a single human gene using GTEx, including full tissue-level expression, publication-quality barplot output, and a simple expression-pattern classification such as universally expressed, tissue-specific, mixed, or non-expressed.

Use when the user wants comprehensive annotation for one gene or a gene list, including gene identifiers, protein information, concise functional summary, domain architecture, Gene Ontology terms, pathway annotations, disease associations, and source-backed metadata.

Use when the user wants to run GSEA from a ranked gene list and generate publication-quality enrichment figures and result tables.

Identify potential research collaborators based on topic overlap. Searches PubMed for researchers publishing on given topics, ranks them by publication output and topic relevance, and generates a ranked collaborator report with their research profiles.

Track recent publications from a specific research lab or PI. Given a PI name and/or institution, fetches their recent papers from PubMed and Europe PMC, summarizes research themes, citation impact, and generates a formatted publication report.

Cluster samples or features in omics data using K-means, hierarchical agglomerative clustering, DBSCAN, or consensus clustering. Evaluates cluster quality, assigns cluster labels, and generates heatmaps, dendrograms, and silhouette plots.

Apply PCA, UMAP, or t-SNE dimensionality reduction to omics data matrices (gene expression, proteomics, methylation, etc.). Projects samples into 2D/3D space, colors by metadata, and generates publication-quality scatter plots with explained variance and loadings.

Train and evaluate machine learning classifiers on omics data (gene expression, mutation scores, copy number, or any numeric feature matrix). Supports Random Forest, Logistic Regression, SVM, and XGBoost. Outputs cross-validation metrics, test set performance, feature importances, ROC curve, confusion matrix, and optional UMAP embedding. Works with any two-class or multi-class label.

Use when the user wants a weekly report of recent high-impact bioinformatics papers from selected journals, ranked by journal prestige, generic bioinformatics relevance, and optional user-interest keywords, with a PDF report and tabular output.

Digest a single scientific paper into a structured summary. Accepts a PubMed ID, DOI, arXiv ID, or local PDF path. Fetches abstract and metadata from PubMed/CrossRef APIs, extracts key findings, methods, results, and clinical implications, and saves a structured markdown report.