Predict chromatin states across 919 epigenomic tracks from FASTA sequences via the Genomic Intelligence API.
npx skills add https://github.com/ClawBio/ClawBio --skill gi-chromatin
Please help me install this Agent Skill. Skill: gi-chromatin Source: https://github.com/ClawBio/ClawBio/tree/main/skills/gi-chromatin Command: npx skills add https://github.com/ClawBio/ClawBio --skill gi-chromatin
Running DeepSEA-style chromatin annotation models locally requires custom PyTorch environments and model weight management. This Skill replaces that setup with a single CLI call to the hosted Genomic Intelligence API, returning per-window predictions across 919 tracks (histone marks, DNase, TF binding) for any DNA sequence.
Run the gi-chromatin demo to predict chromatin states for the bundled chr19 active-promoter FASTA sequence.
High-intent search queries and answers about installing and using this skill.
Provide a single-record FASTA file of 200 to 500,000 bp and run the gi_chromatin.py CLI with --input and --output flags. The skill posts the sequence to the Genomic Intelligence chromatin endpoint and writes a Markdown report plus a full result.json with per-window predictions across 919 tracks.
DeepSEA-style prediction uses a deep learning model to score a DNA sequence across hundreds of epigenomic tracks, including histone modifications, DNase hypersensitivity, and transcription factor binding. This skill calls a hosted 919-track model so no local PyTorch environment or model weights are needed.
Yes, the skill requires a Genomic Intelligence partner key supplied via the --api-key flag or the GI_API_KEY environment variable. A shared hackathon-tier key ships in the repo's .env.example for opt-in demo use, and individual keys can be requested by email.
The API accepts sequences from 200 to 500,000 bp after whitespace stripping, and the skill validates both bounds locally before sending a request. Sequences under the model's 1,000 bp context window are still scored but padded, and the skill warns in that case.
This skill uploads your FASTA sequence to the hosted Genomic Intelligence API, unlike local-first ClawBio skills. You should not submit identifiable patient data without an appropriate data-use agreement, and the bundled demo uses public reference sequence only.
K-Dense-AIRead, query, filter, and write SAM/BAM/CRAM, VCF/BCF, FASTA/FASTQ, and tabix genomic files with pysam.
ammawlaClassify genomic regulatory elements from ENCODE chromatin signatures.
mims-harvardAnalyze histone ChIP-seq, ATAC-seq, chromatin states, and TF binding from ENCODE and GTEx.
thmtzRoute NanoClaw agent groups to a local Ollama model via configuration changes.
