Looking at Prusty's talk I think he is doing sensible experiments. The biggest problem is that his sample sizes are way too low which makes a lot of it hard to believe in my opinion. It's cool I think to look at an antigen array and it is a good follow up to the 'factor in the blood' to isolate antibody and see if that is sufficient to induce the phenotype he's described before of mitochondrial fragmentation:
I'm not sure what to make of the fact he sees higher mitochondrial area in mild and low in severe, but altogether doesn't feel very conclusive to me. I would have also like to have seen if the remaining antibody depleted serum had an effect.
One of the central criticisms of his 2020 immunohorizons paper is that he talks about ME serum and talks about herpes biology but there's no actual link between those two experimental systems so it is spurious to conclude from that work that they are linked. The link he provides this time comes from ELISA for antibodies to the herpes protein dUTPase:
He's arguing there is greater anti-dUTPase antibody in ME and long covid but honestly I don't like inferring that from a percentage based plot like this. Antibody levels have been binned into 0,1,2,3 categories corresponding to low and high levels respectively. I'd like to see a scatter plot and judge for myself!
He looks at IgG and IgM antigen specificities by putting them onto an antigen array. He claims to not see anything in IgG but that the IgM array perfectly separates the healthies, milds and severes:
However the dendrogram on the left actually isn't showing that. It is forming two groups, one containing 2 healthies and the milds, and another containing 1 healthy and the severes. Besides that 3 controls is just not even close to enough to make any reasonable conclusions imo. I would love to see this same experiment but with n=100 in each group.
He is claiming to see higher levels of serum fibronectin:
If each of those dots is a patient then it looks like a good sample size and possibly a subgroup of patients have higher FN1. He's arguing in the rightmost scatter plot that FN1 levels correlate with severity. All this would be good to see replicated somewhere else. Plasma proteomics have been done before in a couple of different contexts and I don't think FN1 has been picked up before so I'm not sure that's encouraging.
