[x] Update source files for updated, QC'd SCElist's (there will be two separate/updated .rdas for DLPFC, since we added an additional sample for this 'revision' batch):
Those are in: /dcl01/lieber/ajaffe/Matt/MNT_thesis/snRNAseq/10x_pilot_FINAL/rdas/revision/all-FACS-n14_preprint_SCEs_processing-QC_MNTMar2021.rda and /dcl01/lieber/ajaffe/Matt/MNT_thesis/snRNAseq/10x_pilot_FINAL/rdas/revision/all-FACS-n10_2021rev_SCEs_processing-QC_MNTMar2021.rda
[x] cbind() the DLPFC samples (largest to smallest; see below for more elaboration), then perform multiBatchNorm()
These will have the .dlpfc suffix
[x] modelGeneVar and just take any positive 'biological components' (defined here) as the 'HVG' space
[x] run fastMNN() to compute corrected PC coordinates, with merge.order as the order from largest to small sample sizes
Take that "corrected"reducedDims entry from the output and save it into the original SCE
(Aside - if we also had NeuN-enriched samples, would merge these into all the 'total fraction'-samples, merged, first. This approach is described/recommended in the batchelor manual (starting here))
[x] Save this new region-specific SCE & its defined 'chosen.hvgs' Boolean vector into a new .rda in the rdas/revision/ Rdata dir
This should basically pretty much exactly reflect the amygdala 'step02' script up until here
Operating in '10x_DLPFC-n3_step02_clust-annot_MNT.R' and referencing '10x_Amyg-n5_step02_clust-annot_MNT.R':
[x] Update source files for updated, QC'd
SCElist's (there will be two separate/updated.rdasfor DLPFC, since we added an additional sample for this 'revision' batch):/dcl01/lieber/ajaffe/Matt/MNT_thesis/snRNAseq/10x_pilot_FINAL/rdas/revision/all-FACS-n14_preprint_SCEs_processing-QC_MNTMar2021.rdaand/dcl01/lieber/ajaffe/Matt/MNT_thesis/snRNAseq/10x_pilot_FINAL/rdas/revision/all-FACS-n10_2021rev_SCEs_processing-QC_MNTMar2021.rda[x]
cbind()the DLPFC samples (largest to smallest; see below for more elaboration), then performmultiBatchNorm().dlpfcsuffix[x]
modelGeneVarand just take any positive 'biological components' (defined here) as the 'HVG' space[x] run
fastMNN()to compute corrected PC coordinates, withmerge.orderas the order from largest to small sample sizes"corrected"reducedDimsentry from the output and save it into the originalSCEbatchelormanual (starting here))[x] Save this new region-specific
SCE& its defined 'chosen.hvgs' Boolean vector into a new.rdain therdas/revision/Rdata dirThis should basically pretty much exactly reflect the amygdala 'step02' script up until here