skills.rest
Estrella-231
Estrella-231COMMUNITY
Explore all skills by this author →
GitHub

geniml

Generate machine learning embeddings for genomic intervals from BED files.

Updated May 24, 2026
Search Tags:#embeddings#bedspace#region2vec#bed files#scatac-seq#universe building#genomic
One-click install
npx skills add https://github.com/Estrella-231/Mathematical_modeling_tongmeng --skill geniml-estrella-231
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: geniml
Source: https://github.com/Estrella-231/Mathematical_modeling_tongmeng/tree/main/.agents/skills/geniml
Command: npx skills add https://github.com/Estrella-231/Mathematical_modeling_tongmeng --skill geniml-estrella-231

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Geniml turns genomic interval data stored as BED files into useful machine-learning embeddings, avoiding manual feature engineering and enabling fast similarity search and clustering across regions and cells.

Core Features & Use Cases

  • Region2Vec for bulk embeddings: Learn unsupervised vectors for genomic regions/region sets from collections of BED files for similarity, clustering, and downstream ML.
  • BEDspace for joint region–metadata embeddings: Train shared embeddings for region sets and associated labels to support cross-modal queries like region→label and label→region.
  • scEmbed for single-cell ATAC-seq: Generate cell embeddings from scATAC-seq peak coordinates for clustering, annotation, and visualization.
  • Consensus Peaks (universes): Build statistically grounded reference peak vocabularies to standardize tokenization across datasets, using CC/CCF/ML/HMM approaches.
  • Utilities for preprocessing, null models, evaluation, and search backends: Improve repeatability via caching, assess embedding quality, generate randomized controls, and power query interfaces.

Quick Start

Use Geniml to embed genomic regions from your BED collection by running the Region2Vec CLI with your token folder, saving the trained model and embeddings to disk.

Frequently Asked Questions about geniml

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

FAQPage Schema
How do I generate machine learning embeddings from BED files for genomic regions?▼

To generate genomic interval embeddings from BED files, you run the Region2Vec CLI with your token folder, saving the trained model and embeddings to disk. This enables fast similarity search and clustering across genomic regions.

What is a BED-based tokenization universe and when do I need one?▼

A BED-based tokenization universe is a statistically grounded reference peak vocabulary that standardizes tokenization across datasets. You need one to ensure consistent tokens when training embeddings for region sets or single-cell ATAC-seq data.

Can I embed single-cell ATAC-seq peak coordinates for clustering and visualization?▼

Yes, you can generate cell embeddings from scATAC-seq peak coordinates using scEmbed. This requires building a reference universe for consistent tokens, then training and encoding to produce embeddings for clustering and visualization.

How do I train joint region and metadata embeddings for cross-modal retrieval?▼

You train shared embeddings for region sets and associated labels using BEDspace. This joint region–metadata embedding supports cross-modal queries, enabling both region-to-label and label-to-region retrieval workflows.

What is the best way to standardize genomic interval tokenization across multiple datasets?▼

The best way to standardize genomic interval tokenization is building consensus peaks, or universes, using CC, CCF, ML, or HMM approaches. These create a statistically grounded reference peak vocabulary for consistent tokens across datasets.

Do I need to preprocess BED files before learning genomic interval embeddings?▼

Yes, BED files require preprocessing before learning genomic interval embeddings. Utilities for preprocessing, null models, and evaluation improve repeatability via caching, assess embedding quality, and generate randomized controls before training.

Related Skills

Explore More Education & Research Skills→
shushuznshushuzn
4

geniml

Build machine learning embeddings from genomic interval BED files.

Community
Advanced
LeonChaoXLeonChaoX
783

geniml

Build machine learning embeddings from genomic interval BED files.

Community
Advanced
arsityarsity

sentence-transformers

Generate embeddings for semantic similarity, clustering, and retrieval with sentence-transformers.

Community
Advanced
Yezez9Yezez9

geniml

Train unsupervised genomic region embeddings from BED files with geniml.

Community
Advanced