$ ls ~/projects
Genomics databases, teaching simulations, lab provenance and compliance tools, and a couple of paid apps.
6 featured · 6 tools & apps · 3 open-source contributions · 2 client engagements
Small VCF databases, one per cohort.
Embedded ClickHouse for variants, embedded DuckDB for annotations, and an MCP natural-language layer. Trio analysis, a GATK3 CombineVariants reimplementation, and native desktop apps. No daemon, no server.
Population genetics, simulated in the browser.
40 interactive modules covering drift, selection, mutation, mating systems, gene flow, linkage and genomics (SFS, sweeps, PCA, GWAS, ROH), each with a narrated step-by-step walkthrough. A modern rebuild of the PopGene.S desktop package. Nothing to install, no data uploaded.
mtDNA haplogroups, from sequence to report.
A Python reimplementation of Haplogrep3: haplogroup classification against PhyloTree 17.2, MITOMAP disease annotation, and population frequencies from 67,000+ GenBank genomes. Research use only.
An electronic lab notebook that writes itself.
Passively captures the commands you run, the files they touch and how they relate, then lets you explore that provenance in a live web UI. A modern take on the BURRITO system. Private beta.
$ strandline ledger verify --since v1.4.0
$ ▌
Code-review evidence for lab software changes.
Self-hosted change review for lab and bioinformatics software: code metadata, downstream impact, SSO, and a hash-chained, tamper-evident audit ledger.
$ pbpaste | context --detect
$ ▌
A clipboard for macOS that knows what you copied.
Detects JSON, JWTs, timestamps, base64 and URLs as you copy them and lets you paste them in a useful shape. Speaks bioinformatics too: HGVS, rsIDs, VCF lines.
ISCN 2024 karyotype validation: parser, rule engine and AST types for the nomenclature used in clinical cytogenetics. Pure Python, zero dependencies, on PyPI.
Python · TypeScript · Cytogenetics
An MCP server that talks directly to NCBI E-utilities and BLAST, with structured provenance logged for every call. Built for validated and clinical work.
Python · MCP · NCBI E-utilities · BLAST
Model Context Protocol server that gives Claude Code access to Alliance of Genome Resources genomics data and tools.
Python · JavaScript · MCP · Genomics
A fine-tuned 7B code-review model: tree-sitter parsing, trained with Apple MLX, served with llama.cpp and FastAPI, GGUF weights on Hugging Face.
Python · MLX · LLM Fine-tuning · tree-sitter
A retro arcade shooter for Apple Watch: the Digital Crown steers, a tap fires. iPhone companion, iCloud sync and Game Center leaderboards.
Swift · SpriteKit · SwiftData · CloudKit · watchOS
Python module for reading Applied Biosystems .ab1 trace files (Sanger sequencing chromatograms).
Python · Sanger Sequencing
REST API over the WormBase Datomic database, serving the WormBase website per widget and powering tools like BLAST.
Contributions to InterMine, the open-source data warehouse for integrating biological data.
The next-generation interface for InterMine genomic data discovery.
Production pipelines for diagnostic labs: variant calling, annotation, reporting, and HL7 delivery to laboratory information systems.
Variant stores for clinical genomics labs, integrated with laboratory information systems.
Client systems aren't public. → ask about similar work
Full stack of tools and technologies across our projects