geniml

What problem does it solve?

This Skill simplifies complex machine learning tasks on genomic interval data (like BED files), enabling researchers to build sophisticated models without deep ML expertise.

Core Features & Use Cases

• Genomic Embeddings: Train Region2Vec or BEDspace models to represent genomic regions and metadata in a low-dimensional space for similarity searches and analysis.
• scATAC-seq Analysis: Use scEmbed to generate cell embeddings from single-cell ATAC-seq data for clustering and cell-type annotation.
• Universe Building: Create standardized reference peak sets (universes) from multiple BED files for consistent analysis.
• Use Case: A researcher has several ChIP-seq BED files and wants to find regions similar to a known enhancer. They can use geniml to train Region2Vec embeddings and then query for similar regions.

Quick Start

Use the geniml skill to train Region2Vec embeddings on BED files located in the 'bed_files/' directory, saving the model to 'model/'.