If you have data from multiple tissues check if one of the copy is consistently expressed across tissues or if there is some variability in paralog expression. It can tell you which one is under a more ubiquitous regulatory regime. Hence functionally ‘more’ important.

More like directed contingency.

Skepticism is fine, but this feels like outright denouncing a field based on a coarse understanding of the process, and is slightly sensationalist.

I’m afraid this is a misrepresentation of what actually goes on behind the scenes in AI assisted drug discovery. There are benchmarks and validations. No one will take the output of a model face value. It’s not the same as ChatGPT. Also, the article is just about using the technology. Nothing else.

From doctors and scientists of course. AI just streamlines the process of discovery, hopefully, to increase efficiency. I mean, that’s the goal. Implementing correctly is a function of the organisation and the people behind it. So it’s as flawed as the people behind it.

Don’t get the cynicism around this. This is something deep leaning would be good at and has been developed for such roles.

What can you do with a phylogenetic data tensor?

Congrats! Neat dataset!

I think Sir David has a substantial part in it.

Such pretty plots!

No matter how persuasive, they’ll never a be a match for reviewer 2. 😂