geniml

Name: geniml
Availability: InStock
Author: robinbarvaag

Train Region2Vec and BEDspace models on genomic interval data.

Updated Jan 10, 2026

Search Tags:#embeddings #machine learning #genomics #single-cell #atac-seq #bed files

Ready to run

npx skills add https://github.com/robinbarvaag/poynt --skill geniml-robinbarvaag

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: geniml
Source: https://github.com/robinbarvaag/poynt/tree/main/.github/skills/geniml
Command: npx skills add https://github.com/robinbarvaag/poynt --skill geniml-robinbarvaag

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill simplifies complex machine learning tasks on genomic interval data (like BED files), enabling advanced analysis without deep ML expertise.

Core Features & Use Cases

Genomic Embeddings: Train models like Region2Vec and BEDspace to represent genomic regions and their associated metadata in a low-dimensional space.
Single-Cell Analysis: Use scEmbed for unsupervised clustering and analysis of single-cell ATAC-seq data.
Universe Building: Create standardized reference peak sets from collections of BED files for consistent analysis.
Use Case: Analyze thousands of ChIP-seq or ATAC-seq peak files to discover patterns, cluster similar regions, or identify cell types from single-cell data.

Quick Start

Use the geniml skill to train a Region2Vec model on the BED files in the 'bed_files/' directory.

Frequently Asked Questions about geniml

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

FAQPage Schema

How do I apply machine learning to genomic interval data from BED files?▼

To apply machine learning to BED files, you can use unsupervised region embedding models like Region2Vec to represent genomic intervals and their metadata in a low-dimensional space for pattern discovery.

What is the best way to cluster single-cell ATAC-seq data without prior labels?▼

For clustering single-cell ATAC-seq data, you can use the scEmbed tool to generate unsupervised embeddings, enabling the identification of cell types and analysis of chromatin accessibility without requiring labeled training data.

How does universe building work for ChIP-seq and ATAC-seq consensus peaks?▼

Universe building creates a standardized reference peak set from collections of BED files, generating consensus peaks that ensure consistent feature spaces across multiple genomic interval analyses.

Can I jointly embed genomic regions and their associated metadata together?▼

Yes, you can train a BEDspace model to jointly embed genomic regions and their associated metadata within a single low-dimensional space, facilitating integrated analysis of sequence features and experimental annotations.

Do I need external dependencies like StarSpace to train Region2Vec embeddings?▼

Training Region2Vec embeddings requires the geniml and scanpy Python packages, and may also require the external StarSpace dependency for generating the low-dimensional genomic region representations.