A scientific theory IS an agglomeration (no pun intended) of analytical concepts that attempts to explain observed phenomenon. What you call it is irrelevant. Importantly, is it true? Does the model make testable predictions?

So it is semantics. You are referring to the mechanics of phase separation and the maintenance of phase separation in out of equilibrium systems. But the basic definition does not necessarily require nor exclude violation of detailed balance.

Is this semantics? A working def. of phase separation is that a substance exists in two spatially separated forms which have different physicochemical properties at some scale. Now all that remains is to demonstrate that such a thing exists ,and if so, than is it relevant to biology.

The order of authors as a measure of contribution is already an archaic system. Academia needs drop papers as performance metrics in general.

You could paint antelope and see if that reduces their chances of being caught.

Pretty cool to have computational biologists and medical omics people side by side. Omics is not systems biology though.

That’s a wonderful initiative. Many don’t realise how to test properly is one thing, and how to come with things to test is the real progress. We are after all limited by our imagination.

I would just plot log2FC(Protein) vs. log2FC(mRNA) and each point is a cell type. That way, everything on the diagonal is obviously seen to be well correlated.

The line interpolating between different tissue/cell types made me cringe. Consider a scatter plot?

It is hardly a 'feel free to' situation - didn't you just mention the funding stream problem in your thread ?

Any other kind of system will be suspected to be susceptible to (and rightly so) forming cliques and mafias of scientists who always accept/promote each others work. That happens even in the current system, but a system that has no check against this is unlikely to work with cynical bureaucrats.

Publications and impact factors are used like that by granting agencies - with confidence in the metric coming from the assumption that it is essentially adversarial peer review refereed by the journal editor.

While I am completely on board with the idea of abandoning the current publication system, any other system is doomed to fail if it does not provide a performance metric *that can be understood by non-scientists*.

Right. I which case the pValue histogram will not show higher frequencies at 0. Or am I understanding your point wrong?

It’s a double edged sword - first, it would be wrong judge people for using AI. AI can automate a lot of dumb tedious tasks. Second, if an AI can generate a particular hypothesis, then it is likely a trivial and obvious thing anyway.

Wait, does this not assume that a ‘real difference’ occurs because of a large number of genes being differential. In reality, there could be a small number of differential genes that are of interest. FDR correction would than control for whether their p-values are really lower than the expected.

This is also very interesting in that the sinusoidal motion of the fins creates lift and thrust simultaneously. I am not an expert this seems different from flying as far as birds and insects go. Air simply does not have the viscosity and weight for this mechanism work at a reasonable frequency

I dread this because as a scientist, Libreoffice extremely limited in certain aspects - Calc is limited to 1024 columns. Impress does not handle videos as seamlessly as Powerpoint, and none of these have good cloud-collaboration and storage. Also, the UI is broken on MacOS.