Reproducible pipelines, RNA-seq workflows, and systems that hold.
I'm Nyssa Ndey-Bongo — a bioinformatics analyst who came into the field the long way, and reads the data better for it.
// Raleigh–Durham–Chapel Hill · remote-friendly
— where the work happened —
$ cat ./about.md
Most bioinformatics pipelines aren't broken. They're just not built to last.
They work — until someone tries to rerun them on a different machine. Until a collaborator inherits the project. Until six months pass and even you can't reconstruct what you did.
I build the other kind. Reproducible RNA-seq and genomics workflows that survive a new laptop, a new teammate, and a future version of you. I spent two years mapping high-dimensional GBS/VCF data to real phenotypes in a genomics lab — and learned that the bottleneck is almost never the data. It's the question, and whether anyone can trust how you answered it.
I came into this field the long way. That route is exactly why I notice the batch effect wearing a biology costume, and why I write pipelines a stakeholder can actually follow. Rigor you can audit; a point of view you can feel.
$ ls ./work
Count matrix → tidy → batch-correct → differential expression → volcano. A diff-expression workflow built to rerun cleanly on any machine, six months later, by anyone.
Mapping phenotype to genotype in high-dimensional blueberry GBS/VCF data — linking traits to the loci that drive them, on genomes most tools quietly assume are diploid.
Reproducible-by-default scaffolding: environments pinned, every step logged, results you can reconstruct. A pipeline is a sentence — if you can't say what it's for in one line, neither can the machine.
Turning genomic signal into something a CFO, a PI, and a future teammate all understand. The best tools I've built started as accommodations for how my own brain works — and tend to help every mind.
$ git log --writing
GitHub
Pipelines, workflows, and builds — in progress and versioned. The source of truth.
Open repo →Codeberg.org
Everything on GitHub, backed up here. The receipts, redundant by design.
Open mirror →Field Notes · LinkedIn
A running series on the craft — what shipped, what broke, and the one decision that saved a week.
Read the series →$ ./start --conversation
Hiring for a genomics or comp-bio team? Need a reproducible RNA-seq or QTL workflow that won't break in three weeks? That's the work I do. Tell me what you're trying to answer.