1. I love this kind of video. Results within days of the experiments being done!! This is the new frontier of science - all researchers should do it like this!! The fact he can give us what seems like a lot and still save enough for publication is amazing. I respect him being honest about the fact he's not giving us the whole picture because of the need to publish.And it is not just good for us - it is good for other researchers.
2. The question of subgroups is always interesting. I'd guess the patient charts he showed were the cleanest clearest examples of each subgroup. There must be edge cases that reduce the tightness of the correlation and were candidates for being chucked in bin number 3. I hope the final paper shows all the participants data not just the means, so we can judge whether the subgroups are truly distinct.
3. The question of the accuracy of the subgroups is especially the case since he is tracking many candidate substances in the blood. If you look at enough analytes, you will find a spurious correlation in at least one due to random chance. So the existence of the groups hypothesised will have to depend on validation.
4. From what I see, Fractalkines are also associated with T cells and endothelial cells (albeit, I admit, reading the wikipedia hasn't enlightened me on the nature of that association...!)
https://en.wikipedia.org/wiki/CX3CL1
"...Soluble CX3CL1 potently chemoattracts
T cells and
monocytes, while the cell-bound chemokine promotes strong adhesion of leukocytes to activated endothelial cells, where it is primarily expressed..."