1/ Claude Code highlighted the wrong gene in my volcano plot. The Ensembl ID mapped to something else entirely, and nothing in the output warned me.

Two tools for plotting genomics data: figeno: Tool for making genomics figures in #python. github.com/CompEpigen/... Plotting Multi-omic Data with plotgardener in #rstats phanstiellab.github.io/plotgardene...

If your bioinformatics analysis isn't reproducible, it isn't finished. Not by you next month, not by a reviewer, not by the next student. Reproducibility is the job.

Just published: "4 tests before you trust vibe-coded results". Check it out and subscribe to get everything I'm making. divingintogeneticsandgenomics.kit.com/posts/4-tes...

4 tests before you trust vibe-coded results

divingintogeneticsandgenomics.kit.com

1/ I’ve reviewed hundreds of bioinformatics GitHub repos in my career. Here’s the brutal truth: most tool documentation fails the people it’s meant to help. And it’s not because the algorithms are bad.

You shared your code. They used it at a conference. No mention. No credit. I've been there. Here's why I still share everything — and why it led to this book.

🧵 You’re not behind in bioinformatics. The truth? The field just evolves faster than any single human can follow.

The person most likely to need your documentation is you, six months from now — staring at your own code with no idea how it works. Documentation isn't optional.

Keith Baggerly calls it “forensic bioinformatics.” His talk proves reproducibility saves lives. The Importance of Reproducible Research in High-Throughput Biology www.youtube.com/watch?v=7gY... clinical trials started using the wrong results. (watch it!)

The Importance of Reproducible Research in High-Throughput Biology

VideoLectures.NetView the talk in context: http://videolectures.net/cancerbioinformatics2010_baggerly_irrh/View the complete Cancer Bioinformatics Workshop, ...

youtube.com

scplotter provides a set of functions to visualize single-cell sequencing data in an easy and efficient way github.com/pwwang/scpl...

1/ Most bioinformatics errors don’t come from bad science. They come from work you can’t reproduce. And when that happens, you don’t just waste time. You risk patients’ lives.

The person most likely to need your documentation is you, six months from now — staring at your own code with no idea how it works. Documentation isn't optional.

Founding readers get the free 15-lesson sampler now + the lowest price in November 👉 divingintogeneticsandgenomics.kit.com/wtdt_bioinf...

What They Don't Teach You — Founding Waitlist

divingintogeneticsandgenomics.kit.com

Local antibody-driven response to anti-PD-1 blockade www.nature.com/articles/s4...

🧵 You’re not behind in bioinformatics. The truth? The field just evolves faster than any single human can follow.

Messy data is why your "quick" analysis takes all afternoon. One principle — tidy data — makes everything downstream faster and far less error-prone.

On October 27, I’ll be joining a panel discussion on how multiomics is changing the way drug discovery teams understand disease biology and evaluate new therapeutic opportunities.

Fc-optimized GITR antibody enhances a CD4 T cell–dendritic cell crosstalk to promote antitumor immunity www.nature.com/articles/s4...

1/ Ever see a stunning bioinformatics figure and think: “How on earth did they make that?” Let’s break it down so you can recreate it.

Running your RNA-seq steps by hand, one script at a time? A workflow language like Snakemake reruns only what changed — and makes your pipeline reproducible by anyone.

Identification of cycling regulatory T cell precursors as conductors of immune escape during breast carcinoma progression www.cell.com/cancer-cell...

Depending on your data type, use the right plot to tell the story. A nice interactive app to determine what plots you need www.data-to-viz.com/

The simplest reproducibility habit almost nobody teaches: date your directories. 2025-09-15_project beats "final_v2_REAL_final" every single time.

really enjoyed all the episodes from Approved podcasts.apple.com/us/podcast/... If you like Acquired, you will like it but for biotech companies.

γδ T cells modulate anti-tumor immunity in small cell lung cancer www.cell.com/cancer-cell...

Want to master bioinformatics data visualization? I almost quit when I first tried ggplot2. But once I “got it,” it changed how I saw data forever. Here’s the hard-won path that makes ggplot2 click. 🧵

If you're doing bioinformatics without Git, you're one accidental overwrite away from losing weeks of work. Here's the version-control setup every bioinformatician should have.

1/ You tested 10,000 genes. 500 are significant (p < 0.05). Exciting? Maybe. But if you used raw p-values, you just got played by randomness. Let’s talk FDR.

1/ What the heck is an "object" (like a Seurat object) in programming? If you're in bioinformatics, this term shows up a lot. Let’s break it down with examples that actually make sense 🧵

You can install every fancy tool in the world. If you can't move around Unix, you'll still be slow. Unix is the one skill that makes every other bioinformatics skill faster.