skills.rest
openai
openaiOFFICIAL
Explore all skills by this author →
WebsiteGitHub

scrna-seq-qc

Process, filter, annotate, and visualize single-cell RNA-seq datasets with auditable QC pipelines.

5.3k|765|Updated Mar 4, 2026
Search Tags:#bioinformatics#single-cell#umap#scrna-seq#cell annotation#doublet detection#Quality Control
One-click install
npx skills add https://github.com/openai/plugins --skill scrna-seq-qc
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scrna-seq-qc
Source: https://github.com/openai/plugins/tree/main/plugins/ngs-analysis/skills/scrna-seq-qc
Command: npx skills add https://github.com/openai/plugins --skill scrna-seq-qc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Single-cell and single-nucleus RNA-seq datasets require dataset-specific quality control, doublet detection, cell annotation, and visualization, but copying fixed thresholds across tissues and species produces unreliable results. This Skill guides the construction of an auditable scRNA-seq QC pipeline that derives thresholds from observed data and preserves provenance at every step.

Core Features & Use Cases

  • Data-Driven QC Filtering: Choose thresholds for detected genes, total UMIs, and mitochondrial fraction from observed distributions, with required scDblFinder doublet detection and ambient-RNA filtering per batch.
  • Reference-Based Cell Annotation: Annotate cells using matched references such as the Allen Brain Cell Atlas via MapMyCells, or fall back to conservative marker-based labeling when no reference exists.
  • Scalable UMAP Visualization: Generate global and per-group UMAPs from scVI latent spaces or batch-aware marker panels, with memory-safe practices for large AnnData objects.
  • Use Case: Given a 10x Genomics output bundle for a multi-batch brain tissue dataset, build a complete QC pipeline that filters low-quality barcodes, calls doublets per capture channel, annotates cell types against a matched atlas, and produces legible UMAP figures plus a full parameter manifest.

Quick Start

Ask the AI to build a scRNA-seq QC pipeline for your dataset, specifying the organism, tissue, and whether the data are whole-cell or single-nucleus.

Frequently Asked Questions about scrna-seq-qc

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

FAQPage Schema
How do I choose QC thresholds for scRNA-seq data?▼

Choose thresholds from the observed distributions of detected genes, total UMIs, and mitochondrial fraction rather than fixed cutoffs. Plot metrics overall and per batch, pick cutoffs from distribution valleys or robust outlier rules like 3 MAD below the median, and save justification plots.

What doublet detection method should I use for single-cell RNA-seq?▼

Use scDblFinder as the required doublet caller, run per batch or capture channel so it sees realistic collision structure. Split very large batches before calling, and preserve doublet scores and class calls in the output metadata.

When should I use scVI versus PCA for scRNA-seq analysis?▼

Prefer standard PCA or Scanpy workflows for smaller datasets with minimal batch structure. Prefer scVI when integrating across batches, donors, or chemistries, or when the dataset is large and noisy enough that a learned latent space improves robustness.

How do I annotate cell types without a matched reference atlas?▼

Cluster the dataset, call coarse compartments first, then identify canonical and cluster-specific markers to refine labels. Leave uncertain clusters as unknown or ambiguous rather than overlabeling, and prefer conservative hierarchical labels over confident fine labels.

Can scRNA-seq QC pipelines handle very large datasets?▼

Yes, by keeping matrices sparse, avoiding whole-object copies of AnnData or Seurat objects, and using views, backed mode, and chunked operations. Run expensive steps like scDblFinder per batch and write checkpoints after major stages.

Why is there no universal mitochondrial cutoff for scRNA-seq QC?▼

Mitochondrial fraction meaning varies by tissue and assay: high fractions mark dying cells in stressed whole-cell data but are less informative in nucleus data. Choose the cutoff from the observed distribution for each dataset and document the justification plot.

Related Skills

Explore More Education & Research Skills→
ya-wayya-way

sc-data-profiler

Profile raw 10x Genomics single-cell matrices to output QC distributions as JSON.

Community
Intermediate
tondevreltondevrel
19

scanpy

Analyze single-cell gene expression data with AnnData and scRNA-seq workflows.

Community
Advanced
TianGzlabTianGzlab
155

spatial-annotate

Annotate cell types in spatial transcriptomics AnnData objects using marker-based scoring, Tangram, scANVI, or CellAssign.

Official
Advanced
stellaromicsstellaromics

bio-single-cell-cell-annotation

Annotate single-cell cell types using reference models and trained classifiers.

Official
Advanced