scvi-lda

Discover latent transcriptional programs in scRNA-seq data using AmortizedLDA.

1|Updated Nov 20, 2025

Search Tags:#single-cell #scvi-tools #topic-modeling #scrna-seq #scvi #amortized-lda

Ready to run

npx skills add https://github.com/tony-zhelonkin/SciAgent-toolkit --skill scvi-lda

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: scvi-lda
Source: https://github.com/tony-zhelonkin/SciAgent-toolkit/tree/main/skills/scvi-lda
Command: npx skills add https://github.com/tony-zhelonkin/SciAgent-toolkit --skill scvi-lda

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AmortizedLDA provides a scalable topic modeling approach to discover latent transcriptional programs across single-cell RNA-seq data, treating cells as documents and genes as words.

Core Features & Use Cases

Discover shared transcriptional programs across cell types.
Identify gene modules and intermediate cell states with interpretable topic space.
Compare topic structures across conditions or batches using scvi-tools.

Quick Start

Remove MT genes from counts, set up anndata with a counts layer, train AmortizedLDA, and retrieve per-cell topic proportions and gene-topic distributions.

Frequently Asked Questions about scvi-lda

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

FAQPage Schema

How do I discover latent transcriptional programs in single-cell RNA-seq data?▼

Latent transcriptional programs in scRNA-seq data are discovered using AmortizedLDA topic modeling, which treats cells as documents and genes as words to identify shared gene modules and intermediate cell states across large datasets.

What is the best way to identify intermediate cell states in scRNA-seq?▼

Intermediate cell states in scRNA-seq are identified by training AmortizedLDA to extract per-cell topic proportions, providing an interpretable topic-based dimensionality reduction space that reveals transitional cellular states.

Do I need to remove mitochondrial genes before topic modeling single-cell RNA-seq counts?▼

Yes, you must remove mitochondrial genes from raw counts before topic modeling scRNA-seq data, and you need to set up a dedicated counts layer in anndata as a prerequisite for configuring and training AmortizedLDA.

Can I compare topic structures across different batches in scRNA-seq?▼

Yes, you can compare topic structures across conditions or batches in scRNA-seq using scvi-tools, which supports AmortizedLDA for scalable unsupervised gene-module discovery and interpretable topic-based comparisons.

How does amortized LDA work for interpretable dimensionality reduction of scRNA-seq data?▼

AmortizedLDA provides interpretable dimensionality reduction for scRNA-seq by treating cells as documents and genes as words, yielding per-cell topic proportions and gene-topic distributions that serve as a transparent latent space.