List of causation hypothesis for follow up after DecodeME

[EndME]

Regarding AI: what happens when the disease doesn’t follow the known rules? I don’t understand how an AI model would be able to find the solution in that case. Maybe others have more insight?

That was what me previous post was about. Arguably you could still get there if you allow for enough exploration, but you'd need some way to know whether the millions of things you end up with are sensible and that currently to me seems well outside of the scope in medicine.

 

[Jonathan Edwards]

I guess we should investigate whether we have actionable results or not, whoever or whatever outputs these hypotheses is irrelevant.

Of course. I am simply pointing out that so far AI has failed to go beyond a rehash of whatever is prominent in media and reviews. And as we know, it is empty vessels that make the most noise.

I have seen some of Dr Unutmaz's comments on AI. They are mostly about what is promised, not what has been delivered as far as I can see. My way of doing science may be messy and Maverick but it got me results repeatedly where others had got nowhere.

 

[Jonathan Edwards]

one wants to predict the in some sense least likely sequence of events that can still be contextually made sense of.

Yes, I think this is something like it. If putting together all we know about the usual rules leads to a putative mechanism it is likely to be wrong because if it followed the rues everyone would have it. One has to hold in the back of one's head a series of leaps of faith into the near impossible and look for a way that would make them stick on the rare occasion they operated.

You also have to be very clear about what you want your hypothesis to predict. My impression is that AI cannot avoid being primed with all the false assumptions that you find in textbook and review explanations of things. The biggest problem with ME/CFS is that people have thought hypotheses should predict all sorts of things that we don't actually have any reason to think occur.

 

[Jonathan Edwards]

If there is persistent T cell activation would we expect to see more gramzyme B and perforin in the blood?

A little while back I posted a query about circulating T cell findings in things like psoriasis. I am not sure one expects to find much. There need be no obvious shifts in population numbers. In fact the bad cells are likely to be absent from circulation, doing bad things elsewhere.

My thought was that you might pick something up by looking at how circulating T cells interact with other cells in short term culture. That may still not get over the problem but if people don't find much in diseases like psoriasis where we know there are misbehaving T cells then finding nothing in ME/CFS does not rule a T cell-mediated problem.