Hi, I'm Jaret 👋

PhD student in Genetics, Bioinformatics & Computational Biology at Virginia Tech. I build open-source tools at the intersection of plant genomics, machine learning, and LLM tooling.

🔧 What I'm building

data-aggregator-mcp — a unified research-data MCP: search & fetch datasets across Zenodo, DataCite, NCBI omics (GEO/SRA), and literature (PubMed/OpenAIRE) behind one normalized model. uvx data-aggregator-mcp

plant-genomics-mcp — 50 genomics tools across 23 backends (Ensembl Plants, Phytozome, UniProt, AlphaFold DB, InterPro, JASPAR, AraGWAS, KEGG, NCBI BLAST, …) with cross-source synthesis.

plantcv-mcp — PlantCV as a measurement instrument: plant trait numbers and the segmentation overlay they were computed from, and it refuses to return numbers when the segmentation is degenerate.

breedsim-mcp — breeding-scheme simulation over AlphaSimR, built on one rule: a single stochastic run is not a result. Returns per-cycle mean, sd and CI, with no flag that collapses it to a point estimate.

jobd — a self-hostable, GPU-aware job broker for your own machines: capability-tagged routing, and the queue exposed natively to agents over MCP.

ghostcite — deterministic, no-LLM detection of ghost citations: catches author/year that don't match the DOI's CrossRef record, plus retraction flags. Ships as a pre-commit hook and a GitHub Action.

🌱 Also working on

Long-tail CLIP for orchid ID (5,000+ species) · StyleGAN2-ADA for Cattleya · RNA-seq differential expression · auxin / ARF genomics.

🧰 Stack

Python · R · PyTorch · MCP / LLM tooling · Bioconductor

📫 Find me

musharna.github.io · Hugging Face · ORCID 0009-0003-4055-5238