What file formats are supported by this skill?

The skill supports .h5ad (AnnData) files and .h5 files from 10X Genomics Cell Ranger outputs.

Can I customize the filtering parameters?

Absolutely. You can adjust MAD thresholds, hard cutoffs, and minimum cell counts via command-line arguments or by using the modular Python functions provided in the skill.

Why use MAD-based filtering instead of fixed thresholds?

MAD-based filtering is more adaptive to the specific distribution of your dataset, preventing the loss of biological signal in heterogeneous samples compared to arbitrary fixed cutoffs.

Does it support different species like human or mouse?

Yes, you can customize gene patterns for mitochondrial, ribosomal, and hemoglobin content to match human (MT-), mouse (mt-), or other species.

Single-Cell RNA-seq QC

Name: Single-Cell RNA-seq QC
Author: anthropics

byanthropics

•

8,000

•

Ciencia de Datos y ML

Automates quality control and filtering for single-cell RNA sequencing data following scverse and scanpy best practices.

This skill provides a comprehensive workflow for assessing and filtering single-cell RNA-seq data in .h5ad and .h5 formats. It automates the calculation of critical metrics—such as count depth, gene detection, and mitochondrial content—and applies robust Median Absolute Deviation (MAD) based outlier detection. Whether using the high-level automated pipeline or modular building blocks, researchers can generate publication-ready visualizations and clean datasets optimized for downstream analysis like clustering and cell-type annotation.

Características Principales

01Automated QC metric calculation for count depth, gene detection, and MT/Ribo content

028,000 GitHub stars

03Customizable species-specific gene pattern matching for human and mouse data

04Robust MAD-based outlier detection for adaptive and accurate cell filtering

05Comprehensive visualization generation including before/after distribution plots

06Support for .h5ad (AnnData) and 10X Genomics .h5 file formats

Casos de Uso

01Cleaning raw 10X Genomics data for downstream bioinformatics pipelines

02Performing exploratory data analysis to assess sample quality and library complexity

03Standardizing QC protocols across multiple scRNA-seq datasets using scverse best practices

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add anthropics/knowledge-work-plugins single-cell-rna-qc

For use in Claude.ai and ChatGPT

Download Skill