omicverse-single-cell-cellvote-consensus

Name: omicverse-single-cell-cellvote-consensus
Availability: InStock
Author: omicverse

Build consensus cell-type labels for AnnData objects with conflicting single-cell annotations.

13|2|Updated Apr 5, 2026

Search Tags:#bioinformatics #consensus #single-cell #anndata #cell annotation #llm arbitration #local majority

Ready to run

npx skills add https://github.com/omicverse/omicverse-skills --skill omicverse-single-cell-cellvote-consensus

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: omicverse-single-cell-cellvote-consensus
Source: https://github.com/omicverse/omicverse-skills/tree/main/src/omicverse_skills/skills/single-cell-cellvote-consensus
Command: npx skills add https://github.com/omicverse/omicverse-skills --skill omicverse-single-cell-cellvote-consensus

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve?

CellVote solves inconsistent single-cell cell-type labeling when multiple annotators disagree, giving you one consensus label per cluster. It is built for AnnData workflows where labels already exist in obs and you need a reliable final annotation without manually comparing every annotator output.

Core Features & Use Cases

Multi-annotator consensus that combines outputs from SCSA, GPT-based annotation, GPTBioInsightor, scMulan, and PopV.
Online arbitration with an OpenAI-compatible model that uses cluster markers and biological context to choose the most plausible label.
Offline deterministic voting with a local-majority fallback for reproducible runs, testing, or environments without API access.
The workflow centers on ov.single.CellVote and its vote method, which writes the final cluster decision back into each cell.
Useful when you need to standardize noisy labels, resolve disagreements across annotators, or reproduce a PBMC3k consensus workflow.

Quick Start

Load your AnnData object, ensure cluster labels and annotator columns are populated, and ask the Skill to produce a consensus CellVote_celltype assignment using either online LLM arbitration or the offline local-majority method.

Frequently Asked Questions about omicverse-single-cell-cellvote-consensus

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

FAQPage Schema

How do I resolve conflicting cell-type labels from multiple single-cell annotation tools?▼

To resolve conflicting cell-type labels from multiple single-cell annotation tools, you can use a consensus voting method that compares cluster markers and biological context to assign one final label per cluster.

Can I run consensus cell annotation offline without an OpenAI API key?▼

You can run consensus cell annotation offline without an OpenAI API key by using a deterministic local-majority fallback that votes based on existing obs label columns for reproducible results.

How do I combine SCSA, PopV, and GPT-based cell-type annotations in an AnnData object?▼

You combine SCSA, PopV, and GPT-based cell-type annotations in an AnnData object by passing populated obs label columns to a cellvote function that arbitrates across the annotator outputs.

What is the best way to standardize noisy single-cell PBMC3k cluster labels?▼

The best way to standardize noisy single-cell PBMC3k cluster labels is applying cluster-level label arbitration that evaluates marker genes and annotator disagreements to write a consensus assignment back into obs.

What inputs do I need for LLM-based single-cell label arbitration?▼

For LLM-based single-cell label arbitration, you need an AnnData object with populated cluster marker genes and existing annotator label columns in obs, plus an OpenAI-compatible model endpoint for voting.

When should I use a local-majority vote instead of online LLM arbitration for cell types?▼

You should use a local-majority vote instead of online LLM arbitration for cell types when you lack API access, need reproducible testing, or want deterministic label results without LLM latency.