omicverse-single-cell-lda-topic-clustering

Run LDA topic modeling on AnnData count data with a MIRA backend.

13|2|Updated Apr 5, 2026
One-click install
npx skills add https://github.com/Starlitnightly/omicverse-skills --skill omicverse-single-cell-lda-topic-clustering
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: omicverse-single-cell-lda-topic-clustering
Source: https://github.com/Starlitnightly/omicverse-skills/tree/main/src/omicverse_skills/skills/single-cell-lda-topic-clustering
Command: npx skills add https://github.com/Starlitnightly/omicverse-skills --skill omicverse-single-cell-lda-topic-clustering

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anndata, numpy, pandas, omicverse, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This skill transforms a notebook-style LDA topic modeling workflow into a reusable, triggerable job that outputs per-cell topic usage and optional hard-label classifications.

Core Features & Use Cases

  • Reusable wrapper ov.utils.LDA_topic that can run in-memory or on-disk.
  • Supports deriving hard cluster labels or RFC-based labels from topic assignments.
  • Works with count-like AnnData data in a modular, auditable workflow for single-cell experiments.

Quick Start

Run a simple predictive workflow by instantiating the LDA_topic wrapper with an AnnData object and calling predicted(num_topics=6) to generate LDA_cluster.

Frequently Asked Questions about omicverse-single-cell-lda-topic-clustering

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

FAQPage Schema
How do I run LDA topic modeling on single-cell RNA data?

To run LDA topic modeling on single-cell RNA data, use the ov.utils.LDA_topic wrapper on an AnnData object containing a counts layer to generate per-cell topic usage predictions.

Can I derive hard cluster labels from LDA topic assignments in AnnData?

Yes, you can derive hard cluster labels or RFC-based labels from LDA topic assignments by applying the LDA_topic.get_results_rfc interface to your AnnData object after topic modeling.

Does this LDA topic clustering workflow support on-disk processing for large single-cell datasets?

Yes, the LDA topic clustering workflow supports both in-memory and on-disk processing modes, which you can toggle using the ondisk parameter for large single-cell datasets.

What is the best way to predict cell topics using a MIRA backend in Python?

The best way to predict cell topics using a MIRA backend is instantiating the LDA_topic wrapper with an AnnData object and calling predicted(num_topics=6) to generate LDA_cluster labels.

Do I need a specific counts layer in my AnnData object to perform topic modeling?

Yes, you need count-like single-cell data within an AnnData object that includes a counts layer, specified via the layers parameter, to properly execute the LDA topic modeling workflow.

What parameters control the number of topics and threshold in single-cell LDA clustering?

The num_topics parameter controls the number of generated topics, while the LDA_threshold parameter sets the assignment threshold for deriving hard cluster labels in single-cell LDA clustering.