bulk-deg-analysis_skill

This skill guides you through bulk RNA-seq differential expression analysis in omicverse, from gene ID mapping to visualization and pathway enrichment.

Python

866

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill starlitnightly/omicverse --skill bulk-deg-analysis

reference.md1.6 KB
SKILL.md5.4 KB

Overview

This skill guides Claude through omicverse's end-to-end bulk RNA-seq differential expression (DEG) pipeline, from raw gene-level counts to normalized statistics, visualization, and pathway enrichment. It assumes you have featureCounts-style count matrices and want reproducible DEG results inside omicverse. The instructions cover ID mapping, DESeq2-style normalization, statistical testing options, plotting, and enrichment workflows.

How this skill works

The skill walks through preparing mapping assets, loading and cleaning raw counts, and converting gene identifiers to symbols. It shows how to build a pyDEG object, normalize using DESeq2 size factors, run DEG tests (Welch t-test by default, or edgepy/limma alternatives), filter results, and generate volcano and per-gene boxplots. Optional pathway enrichment and multi-library visualization steps use downloadable gene sets and built-in plotting functions.

When to use it

You have raw gene-level count matrices (featureCounts) and need DEG analysis.
You must convert Ensembl/Ensembl-like IDs to gene symbols before analysis.
You want DESeq2-style normalization but flexible testing (ttest, edgepy, limma).
You need publication-ready volcano plots and per-gene boxplots from omicverse.
You want to run pathway enrichment on DEG results using curated libraries.

Best practices

Download and store gene ID mapping pairs under genesets/ before analysis to ensure reproducible mappings.
Clean sample column names (strip .bam suffixes) so group labels exactly match counts columns.
Drop duplicate gene symbols keeping the highest expressed entry to avoid ambiguous rows.
Run dds.normalize() to compute size factors prior to any differential testing.
Filter lowly expressed genes (e.g., log2(BaseMean) > 1) before setting fold-change and p-value thresholds.
Save dds.result and enrichment tables to CSV and export figures with plt.savefig() when running non-interactively.

Example use cases

Normalize a mouse tumour featureCounts matrix with DESeq2, run t-test DEG, and plot the top 8 genes on a volcano plot.
Convert Ensembl IDs to symbols, run limma-style modelling for a complex design, and plot boxplots for selected marker genes.
Run edgepy-style DEG between treated and control replicates and perform GO/pathway enrichment on significant genes.
Download pathway databases, prepare genesets for Human or Mouse, and generate combined multi-ontology enrichment plots.

FAQ

Prebuilt mappings include T2T-CHM13, GRCh38, GRCh37, GRCm39, danRer7, and danRer11; you can also generate mappings from local GTF files.

Which DEG methods can I use?

Default is Welch t-test (method='ttest'). Alternatives are method='edgepy' for edgeR-like tests and method='limma' for limma-style modelling.

Do I need internet access to run enrichment?

Internet is required the first time to download gene mappings or pathway databases; subsequent runs can use local copies. Provide a background gene list if internet is unavailable.