analysis-geniml

To train region embeddings for BED files, use the Region2Vec or BEDspace models to represent genomic intervals and their associated metadata in a low-dimensional space. This Skill automates the Python-based feature learning workflow.

The best way to cluster single-cell ATAC-seq data is by using the scEmbed model to train embeddings, which can then be utilized with scanpy for downstream dimensionality reduction, clustering, and visualization.

Yes, you can build a consensus peak set, referred to as a universe, from multiple BED files. This Skill provides functionality to construct these universes for standardized genomic analysis across collections.

No, custom scripting is not required. This Skill streamlines machine learning tasks on genomic interval data, automating complex dimensionality reduction and feature learning workflows that would otherwise need extensive bioinformatics expertise.

Region2Vec trains embeddings for BED file collections, while scEmbed specifically targets single-cell ATAC-seq data for clustering and analyzing distinct cell populations. Both represent genomic regions in a low-dimensional space.

You should avoid using this Skill if your scATAC-seq analysis requires non-Python-based workflows or falls outside standard genomic interval machine learning tasks like region embeddings, clustering, and consensus peak set building.