You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
(1) Regarding visualization with the DoHeatmap function, I've noticed that it uses data from the scale.data slot, which only contains variable genes. However, not all of the genes on my list are included in the variable genes. As a result, these genes are missing from my heatmap.
Would it be appropriate to rescale the SCT assay data (using the ScaleData function) before using the DoHeatmap function to include these non-variable genes in the heatmap?
(2) Although I specified 5000 genes as integration features,
the integration object finally contains scale.data with 5135 genes, which does not match the var.feature count that I had specified.
Could someone please clarify if there is an error on my part or if there's a misunderstanding?
(3) For additional analysis, such as running the AverageExpression function and conducting differential expression analysis with the FindMarkers function, should I specify the assay as RNA or SCT? Additionally, which slot data should I use for these functions?
Thank you,
BJ
The text was updated successfully, but these errors were encountered:
(1) To address this, when running SCTransform, you can set return.only.var.genes = F, so that you can also compute Pearson residuals for genes beyond the variable features.
(2) We will investigate this behavior!
(3) To clarify, if you are wanting to run pseudobulk-based DE via AggregateExpression followed by FindMarkers, you should be using the counts slot and set the test in FindMarkers to DESeq2.
I've recently started merging 36 datasets using SCTransform and removing batch effects using harmony.
Here's what I have been up to:
However, a few points remain unclear to me:
(1) Regarding visualization with the
DoHeatmap
function, I've noticed that it uses data from thescale.data
slot, which only contains variable genes. However, not all of the genes on my list are included in the variable genes. As a result, these genes are missing from my heatmap.Would it be appropriate to rescale the
SCT
assay data (using theScaleData
function) before using theDoHeatmap
function to include these non-variable genes in the heatmap?(2) Although I specified 5000 genes as integration features,
the integration object finally contains
scale.data
with 5135 genes, which does not match thevar.feature
count that I had specified.Could someone please clarify if there is an error on my part or if there's a misunderstanding?
(3) For additional analysis, such as running the
AverageExpression
function and conducting differential expression analysis with theFindMarkers
function, should I specify the assay asRNA
orSCT
? Additionally, whichslot data
should I use for these functions?Thank you,
BJ
The text was updated successfully, but these errors were encountered: