geniml

Yes, you can cluster single-cell ATAC-seq datasets using scEmbed for unsupervised learning and cell annotation. It processes genomic coordinates from anndata objects to integrate single-cell chromatin accessibility data into a unified embedding space for cross-study cell-type discovery.

The best way to build a consensus peak universe from multiple ATAC-seq datasets is using a statistically rigorous framework to merge genomic intervals. This constructs a unified peak set across disparate experimental datasets, facilitating cross-study region embedding and cell-type annotation.

Yes, you need PyTorch and Scanpy to process single-cell ATAC-seq data, along with anndata, pandas, and scipy. These dependencies provide the foundational environment for Python-based processing of genomic coordinates and integration with standard bioinformatics workflows.

Genomic embeddings for cross-study cell type annotation work by training unsupervised models to map genomic regions and cells into a shared feature vector space. This allows researchers to integrate disparate single-cell ATAC-seq datasets and transfer annotations across different experimental conditions.

A limitation when integrating disparate ATAC-seq datasets into a unified embedding space is the requirement for Python-based processing of genomic coordinates using standard bioinformatics data structures. Incompatible BED file formats or missing metadata can hinder unsupervised region embedding and cross-study annotation.