geniml

To convert BED files into genomic embeddings, you can use this skill to transform raw genomic interval data into structured vector representations for similarity searches and clustering. It processes BED files to build embedding spaces suitable for comparative genomic analysis.

Yes, generating joint embeddings for genomic regions and associated biological metadata such as cell types or tissues is supported. This creates a shared embedding space for simultaneously analyzing relationships between genomic intervals and their biological context.

Single-cell ATAC-seq workflows are supported for chromatin accessibility analysis, aiding in cell type annotation and clustering. It processes single-cell genomic data to help identify distinct cell populations based on chromatin accessibility patterns.

Building a consensus peak universe from multiple datasets involves deriving unified peak sets to standardize genomic regions for comparative analysis. This skill constructs universal peak sets, standardizing regions across different datasets for comparative genomic analysis.

Processing genomic interval data requires specific libraries including scipy, pandas, scikit-learn, h5py, and starspace for machine learning and statistical evaluation. These dependencies support genomic data processing, embedding generation, and model quality evaluation.

Available utility tools include caching, randomization, evaluation, and search utilities designed to streamline genomic research workflows. These utilities help manage data preprocessing, evaluate resulting model quality, and optimize embedding searches across bulk and single-cell datasets.