Hello Bioinformatics lovers,

Tommy here. I was interviewed yesterday by Nature yesterday, about best practices using

spreadsheets and naming files.

P.S.: This is a must read for any scientists (both wet and dry) https://www.tandfonline.com/doi/full/10.1080/00031305.2017.1375989​

It turns out to be a big problem for bioinformaticians.

How many hours do bioinformaticians lose matching sample IDs across assays?​
Too many.
And every one of them is avoidable.

Let’s talk about why this keeps happening—and how we can stop it.

The Chaos Begins with a Name

The same cell line, labeled differently:

• MCF7
• MCF-7
• MDA-MB-361
• MDA_MB_361
• MDAMB361

And that’s within the same team.

Then Come the Timepoints

Add time, treatment, or dose… and now you’ve got:

• MCF7_24h
• MCF7_48h
• MCF7_resistant
• MCF7_combo_dose2

And this is just one dataset.

The Multi-Omics Nightmare

Team A runs RNA-seq.
Team B runs proteomics.
Each with their own naming system.
No coordination. No UUIDs.

Then someone says:
​“Let’s integrate the data!”​
You sigh—and open a spreadsheet that will ruin your evening.

Sound Familiar?

You spend 3 hours cleaning it.
You're still unsure if MCF7_24h_A is the same as MCF-7_T24.
Then you find out…
The same sample was sequenced twice, under two different names, in the same assay.

No one noticed.
Resources wasted.
Trust gone.

This Isn’t Just a Small Lab Problem

Even large-scale efforts like TCGA, pharma pipelines, and multi-million dollar projects suffer from mislabeling and duplication.

Why?
Because data pipelines are only as good as the names they start with.

The Fix Is Boring—but Powerful

• Assign a UUID (universally unique identifier) to every biospecimen
• Centralized sample registry
• Enforced naming conventions
• Robust sample tracking
• Communication between wet lab and data teams

This isn’t rocket science.
It’s just science done right.

In a Small Company?

You can fix this in a day.

In a Large Company?

It’s harder.
Siloed teams. Legacy systems. lack communication

But fixing it will save millions per year.
And preserve trust in your data.

Final Thought:

Bioinformatics isn’t janitorial work.
It’s science.
But right now, we're wasting time cleaning up names like MCF7, MCF-7, and MCF_7.

That’s not where our talent belongs.

Takeaways:

• Bad naming = wasted hours and broken analyses
• Use UUIDs to align assays across teams
• Communication is infrastructure
• Let bioinformaticians do real science, not name-guessing

If you're spending time cleaning up naming errors, it's not your fault.
But it is time to demand better.

Start with one rule:
​Every sample gets a universal name.

Other posts that you may find useful

1. chatomics! New youtube video: How to calculate partial correlation controlling cancer types
2. Deep neural networks (DNN) are just glorified linear models?
3. ggfortify::autoplot() for your PCA plot may not be what you want, here is the details​
4. I am never a handyman, but when your wife mandates you to repair the screen, you learn how to do it. Here is how you can learn anything, including bioinformatics.
5. Here is how I would start learning bioinformatics: A roadmap.
6. How to organize your bioinformatics project without chaos.​
7. Bioinformatics is hard because of decisions. Thousands of tiny choices that shape your results. which CpG, reference, isoform.
8. ​Sample swaps are the silent killer in bioinformatics.
9. Ever wondered why analyzing RNA-seq data feels like walking through a fog with 20,000 dimensions? ​
10. You can not do Bioinformatics with hard cutoffs and thresholds. what's your log2FC cutoff?
11. You’re analyzing 10x Genomics single-cell RNA-seq and notice lots of intronic reads. Let’s unpack why introns show up—and why they matter. 🧵

PPS:

If you want to learn Bioinformatics, there are four 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!