Hello Bioinformatics lovers,

Happy July 4th! If you celebrate it. We had two days off, so I enjoyed staying at home reading books.

The best way to learn something new is by reading.

Then apply what you read to a practical problem.

Repeat, practice.

You will find a better version of yourself in 10 years!

I had a conversation on a podcast talking about AI and bioinformatics. Hope you enjoy it!

Today, we will talk about single-cell studies. How many cells are needed?

Another single-cell study drops.
500,000 cells sequenced.
More UMAP plots. More clusters.

But here’s the question:
​Are we actually learning more—or just counting better?

Single-cell technology was a game-changer.
It opened the door to rare cell types, subtle states, and unseen biology.
No one’s debating that.

What is worth debating is what we’re actually doing with this power.

Most headlines?

Sure. But that’s not new.
Pathologists knew that decades ago—long before RNA sequencing was cool.

We’re stuck in a cell-counting arms race:
10,000...
100,000...
A million cells.

But after a certain point, more cells ≠ better insight.
​It just means more noise—unless you're asking better questions.

Too many studies stay in “descriptive mode”:

• UMAPs and violin plots
• Marker genes and dot plots
• A sea of clusters with no biological anchor

It looks impressive—until you ask what it all means.

The real prize?​
Not clusters.
But causality.

• Regulatory rewiring
• Therapy resistance states
• Cell fate trajectories that change what we do in the clinic

The best single-cell work doesn’t just map cell types.
It connects:

• Single-cell + spatial + proteomics
• Genomics + clinical outcomes
• Data + hypothesis

Let’s stop equating "more cells" with "better science."​
Start asking:

• Did this study change how we think?
• Did it generate a testable hypothesis?
• Does it push biology forward?

The next frontier isn’t deeper sequencing.​
It’s deeper thinking.

Not just more resolution—more relevance.
Not just clustering—causality.

So the next time you read a single-cell study, ask:

• What’s the biological insight?
• Did they connect the dots—or just color them?

Because the story isn't that tumors are complex.
It’s what we do with that complexity that matters.

If this resonates, share it with a colleague who's swimming in UMAP plots.​
Let's raise the bar—together.

Other posts that you may find useful:

1. Use public data before generating your own. why?​
2. ​BiocPy: Facilitate Bioconductor Workflows in Python
3. In bioinformatics, knowing which tool to use is as valuable as building from scratch. Here's why 🧵
4. ​GeneRanger is a web-server application that provides access to processed data about the expression of human genes and proteins across human cell types, tissues, and cell lines from several atlases.
5. chatomics! From Salmon to DESeq2: RNAseq Data Analysis ​
6. How is your p-value histogram look like? blog post: Downstream of bulk RNAseq: read in salmon output using tximport and then DESeq2 ​
7. Do a comprehensive QC before any analysis.​
8. ​What is a (Seurat) object in OOP. ​
9. You think RNA equals protein? Not always.​

PS:

If you want to learn Bioinformatics, there are other ways that I can help:

1. My free YouTube Chatomics channel, make sure you subscribe to it.
2. I have many resources collected on my github here.
3. I have been writing blog posts for over 10 years https://divingintogeneticsandgenomics.com/​

Stay awesome!