scArches FAQs

Question 1

Does it support multi-modal data analysis?

Accepted Answer

Yes, the skill provides technical guidance for multi-modal integration, including transcriptomics combined with chromatin accessibility, protein data (totalVI), and TCR data (mvTCR).

Question 2

What does the scArches Claude Code Skill do?

Accepted Answer

This skill enhances Claude with specialized knowledge for the scArches library, enabling it to assist with deep learning-based single-cell integration, model surgery, and mapping query data to reference atlases like the Human Lung Cell Atlas.

Question 3

What is 'model surgery' and how does Claude help?

Accepted Answer

Model surgery is a technique to adapt pre-trained models to new query data without retraining from scratch. This skill provides the specific code structures needed to load references, freeze embeddings, and add query-specific parameters efficiently.

Question 4

When should I use this skill?

Accepted Answer

Trigger this skill when you need to perform scATAC-seq preprocessing, integrate multi-omics datasets, transfer cell-type labels, or implement complex architectures like scVI, SCANVI, TRVAE, and totalVI.

Question 5

How does this skill improve my bioinformatics workflow?

Accepted Answer

It provides instant access to optimized implementation patterns, recommended hyperparameters, and troubleshooting guidance for model training, significantly reducing the time spent on manual debugging and documentation lookup.

scArches

Key Features

Use Cases

scArches

Key Features

Use Cases