Software

I develop and contribute to open-source tools and skills for genomic analysis, clinical AI, and nutrigenomics. My work spans standalone repositories on GitHub and contributions to community platforms in the bioinformatics AI ecosystem.

Areas of work

Clinical decision support & machine learning

Models and pipelines developed in the context of the Neotree project and related work in clinical AI. Includes data preprocessing, model training, validation frameworks, and performance evaluation in low-resource settings.

Genomics & bioinformatics

Tools for working with genetic and sequencing data — population structure analysis, molecular evolution metrics, and integration with standard bioinformatics formats (FASTA, NEXUS, VCF).

Community contributions

ClawBio — dnasp skill

ClawBio is the first bioinformatics-native AI agent skill library — local-first, reproducible, and built on OpenClaw. I contributed the dnasp skill: a Python reimplementation of DnaSP 6 for population genetics, covering 16 analyses including nucleotide diversity (Pi, Theta-W), Tajima's D, Fu & Li D*/F*, Ka/Ks, McDonald-Kreitman test, Fst, site frequency spectrum, and codon usage bias. Accepts FASTA or NEXUS input; 355 tests.

ClawHub — nutrigenomics skill

ClawHub is the public skill registry for the OpenClaw agent platform. I published and maintain the nutrigenomics skill (v0.3.1): a local, privacy-first tool that generates a personalised nutrition report from consumer genetic data (23andMe, AncestryDNA, or VCF). It analyses 40+ genes and 58 SNPs covering macronutrient metabolism, micronutrient absorption, omega-3 pathways, caffeine, and food sensitivities — with a radar chart, gene × nutrient heatmap, and full reproducibility bundle.

Languages & tools

Python

pandas · NumPy · scikit-learn · statsmodels · lifelines · matplotlib · seaborn

Reproducible research

Git · GitHub Actions · Jupyter · Docker · SHA-256 checksums · OpenClaw reproducibility bundles

Contribute or collaborate

If you find any of my tools useful, have suggestions, or would like to collaborate, please open an issue on the relevant repository or get in touch directly.