How do you clean GEO metadata before downstream analysis?

Question

[Alba]

Hey everyone,

I’ve been working a lot with GEO datasets lately and I always find the metadata messy and inconsistent (organism names vary, missing values, etc.). Especially when you want to combine multiple datasets or do a meta-analysis, it becomes a real bottleneck.

I’m curious:

• How do you normally clean the metadata?
• Do you use your own scripts?
• Do you trust the platform annotations?

I recently stumbled upon a small browser-based tool that lets you input a GEO ID or upload a .soft file and it gives back cleaned metadata (normalizes organism names, highlights missing values, etc.).

Here’s the tool in case you want to take a look:
metagenclean[dot]streamlit[dot]app (Works best in Chrome; Safari has some JS issues)

I’d love to hear your thoughts:

• Would this be useful in your workflow?
• Anything missing that you’d want it to do?
• Any red flags?

Really appreciate any feedback!

[yura.grabovska]

Again... "I recently stumbled upon a small browser-based tool"

On the site: "Created by alba0gf"

That's the second time you've linked your own tool under the pretence of "this is just something I found online"

[jared.andrews07]

Agreed, just make a "Tool" post if you want to spread the word about it.

People tend to take those much better than these weak attempts at nesting a recommendation in a pointless question.

edit: Oh, I just saw, you did make a tool post as well. In that case, this question should be removed.

[yura.grabovska]

They did make a Tool post originally, but still phrased it as something they just came across rather than "I made thing". I dunno, it's odd.