skills.rest
cpanse
cpanseCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

scevan-analysis

Identifies malignant cells in scRNA-seq data via CNV analysis with SCEVAN on FGCZ infrastructure.

Updated May 28, 2026
Search Tags:#bioinformatics#seurat#scrna-seq#cnv#scevan#tumor classification#subclones
One-click install
npx skills add https://github.com/cpanse/skills --skill scevan-analysis-cpanse
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scevan-analysis
Source: https://github.com/cpanse/skills/tree/main/single-cell-spatial-general/skills/scevan-analysis
Command: npx skills add https://github.com/cpanse/skills --skill scevan-analysis-cpanse

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires SCEVAN, Seurat.

What problem does it solve? Distinguishing tumor cells from normal cells in single-cell RNA-seq data is difficult without genomic evidence. This Skill runs SCEVAN copy number variation analysis on FGCZ HPC infrastructure to classify malignant cells, detect subclones, and extract per-chromosome CNV profiles with correct output parsing. ## Core Features & Use Cases - CNV-based tumor classification: Runs SCEVAN pipelineCNA with carefully selected diploid reference cells to label cells as tumor or normal. - Subclone and per-chromosome CNV extraction: Parses SCEVAN output RData files (CNAmtx, CNAmtxSubclones, count_mtx_annot) and maps gene symbols to chromosomes for per-chromosome CNV scores. - HPC-ready workflow: Provides SBATCH templates with memory guidance scaled to dataset size, plus fixes for common pitfalls like wrong output directories and Ensembl ID mismatches. - Use Case: A cancer researcher with a Seurat object of 17,000 cells uses this Skill to run SCEVAN on the FGCZ cluster, annotate tumor subclones, and generate UMAP visualizations with CNV heatmaps. ## Quick Start Ask the agent to run SCEVAN CNV analysis on your Seurat object using T cells as the reference population and extract tumor classifications and per-chromosome CNV scores.

Frequently Asked Questions about scevan-analysis

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I identify tumor cells in scRNA-seq data with SCEVAN?▼

Run SCEVAN's pipelineCNA function on the raw count matrix from your Seurat object, providing confident normal cells such as T cells as the reference population. The result classifies each cell as tumor or normal and optionally assigns tumor subclones.

What reference cells should I use for SCEVAN CNV analysis?▼

Use confident diploid normal cells, typically immune cells like Memory CD8+ T cells or CD8+ NKT-like cells. Aim for 500-2000 reference cells and avoid populations potentially contaminated with tumor cells.

Why are all cells showing NA for SCEVAN tumor classification?▼

SCEVAN writes results to an output/ subfolder within the working directory, so looking in the wrong path yields no data. Load the CNAmtx and CNAmtxSubclones RData files from that output folder to recover classifications.

Why does per-chromosome CNV extraction fail in SCEVAN?▼

The annotation file uses Ensembl IDs as rownames while the CNV matrix uses gene symbols, so direct matching fails. Use the gene_name column of count_mtx_annot to build the gene-to-chromosome mapping instead.

How much memory does SCEVAN need for large scRNA-seq datasets?▼

SCEVAN is memory-intensive: roughly 17k cells need 512GB (32G per CPU with 16 cores), 10k cells need 256GB, and 5k cells need 128GB. Increase --mem-per-cpu in the SBATCH script if out-of-memory errors occur.

Related Skills

Explore More Education & Research Skills→
stellaromicsstellaromics

bio-workflows-scrnaseq-pipeline

Process 10X Genomics single-cell RNA-seq data into annotated cell types with Seurat.

Official
Advanced
felixboehmfelixboehm

cbioportal-database

Query cBioPortal REST API for cancer genomics data and survival correlations.

Community
Advanced
chenyhvvvvchenyhvvvv
3

cnv-inference

Infer CNV scores and clone labels from spatial transcriptomics AnnData objects.

Community
Advanced
519lab519lab
2

llm-wiki

Automate ingestion and synthesis of markdown-based knowledge base entries.

Official
Intermediate