Builds machine learning models and unsupervised embeddings from genomic interval data and BED files.
Geniml is a comprehensive toolset designed for performing machine learning on genomic regions, enabling researchers to transform raw BED files into actionable embeddings. It provides specialized methods like Region2Vec for unsupervised learning, BEDspace for joint region-metadata analysis, and scEmbed for single-cell ATAC-seq data analysis, making it an essential skill for large-scale genomic similarity searches, clustering, and feature engineering in bioinformatics workflows. This skill guides the implementation of statistically rigorous consensus peak sets and provides standardized patterns for embedding evaluation.
Características Principales
01BEDspace joint region and metadata embedding for search and discovery
02scEmbed optimized for single-cell chromatin accessibility analysis
03Integrated utilities for BED file caching, randomization, and quality evaluation
04Region2Vec unsupervised genomic region embedding generation
05Statistical universe building for consensus peak set creation
060 GitHub stars
Casos de Uso
01Building searchable genomic databases connecting regions with experimental metadata
02Clustering and cell-type annotation for single-cell chromatin accessibility data
03Training high-dimensional feature vectors for genomic region similarity analysis
