Converts genomic interval data from BED files into trainable embeddings for machine learning.
npx skills add https://github.com/Hung-3008/agusta --skill geniml-hung-3008
Please help me install this Agent Skill. Skill: geniml Source: https://github.com/Hung-3008/agusta/tree/main/.agents/skills/geniml Command: npx skills add https://github.com/Hung-3008/agusta --skill geniml-hung-3008
Geniml provides a Python toolkit for building machine learning models on genomic interval data from BED files, enabling unsupervised embeddings, metadata-aware analyses, and universe-based tokenization to drive downstream ML tasks.
Install geniml, prepare BED files with a universe reference, tokenize regions, train a Region2Vec model, and evaluate embeddings.
High-intent search queries and answers about installing and using this skill.
To create genomic interval embeddings, you first need a universe reference for tokenization, then use a Python-based toolkit to train models like Region2Vec on your BED file collections for downstream similarity and clustering tasks.
A genomic universe is a reference set of consensus peaks required for tokenizing genomic interval data. You need it to standardize regions across datasets for joint analyses and single-cell ATAC-seq embeddings.
Yes, the scEmbed feature allows you to train unsupervised embeddings for single-cell chromatin accessibility data, enabling cell-level clustering and annotation for single-cell ATAC-seq datasets.
Yes, a universe reference is required for tokenization. It provides the standardized consensus peaks necessary to convert BED file genomic intervals into trainable tokens for downstream machine learning tasks.
The best way to perform metadata-aware search on genomic regions is by using BEDspace to generate joint region and metadata embeddings, enabling cross-modal queries across your genomic datasets.
A key limitation is the strict requirement for a universe reference; without consensus peaks for tokenization, you cannot standardize genomic intervals for joint cross-dataset analyses or train unified machine learning models.
jackspaceGenerate unsupervised embeddings for genomic intervals from BED files.
LeonChaoXBuild machine learning embeddings from genomic interval BED files.
wsxwj123Learn genomic interval embeddings from BED files using Region2Vec, BEDspace, scEmbed, and Universe.
swaruplabBuild unsupervised embeddings and tokenization workflows for genomic interval data.
