Closed lcolladotor closed 2 years ago
We could also consider separating by WM vs grey matter if needed.
Maybe drop more lowly expressed genes at the pseudo-bulk level?
See filterByExpr()
at https://bioconductor.org/books/release/OSCA/multi-sample-comparisons.html#differential-expression-between-conditions. This would be relevant if we subset to say Layer1, then do DE across pathology levels in just Layer1.
Could be superficial vs deep GM using PCP4 as the boundary for superficial vs deep. Then it would be:
Kristen @kmaynard12: there are no plaques on the WM. Good to know!
Pseudo-bulk at the DLPFC layer (from #9) + pathology resolutions and check if we have enough spots. We can then do a PCA or MDS plot and some other QC.
Pathology resolution levels: