This was specifically aimed at brain transcriptomics:
- using horrendous amounts of data manipulation and 'transformation' and 'cleaning' that would make a statistician cry from relatively small data sets
- to visualize already miniscule differences in epigenetic markers, which are notoriously finicky and hard to correlate to things, given they're shared even across species and different across different cell types (obviously)
- in order to make claims about genetic and epigenetic changes in mood disorders and similar mental illnesses, which are notoriously hard to align with actual specific brain effects that isn't related to any other mood disorder/mental illness, because as far as we know it's all a big cluster of socio-dev issues and may not be concretely affiliated to brain region genetic activity at all
- because the difference between transcriptomics and higher order cognition is equivalent the difference between machine code/hardware level programming and actual AI
yes fmri is also a good pot shot to take. the
salmon study was good for it
I am also a pattern observer. However, there's a lot of layers of potential systemic failure of certain behaviors before you hit "pick a gene and try to blame it". we've got genetic networks, epigenetics, intercellular interaction networks, intracellular interaction networks, local brain systems, local network interactions, global interactions, whatever the fuck is going on with brain waves and mode networks, etc. If you like networks it's, for lack of a better term, a turkduken of like 10 network levels, each massive & beautiful and contributing both up and downstream.
That we've gotten to the point that we can manipulate/edit most of these things is fantastic news though. Shame China will get to it first since they don't hold themselves to ethics.