switch to assertions here too. and clarify Stdev returning abs values…

…. other small fixes.

rmd/31_SCTransform.Rmd

@@ -101,19 +101,25 @@ We're going to use the SCTransform function with some more recent implementation
 library(sctransform)
 library(glmGamPoi)
 ctrl <- SCTransform(ctrl, method = "glmGamPoi", vst.flavor = "v2", verbose = FALSE)
+```
+
+We could've also regressed out covariates during this process.
+
+```{r, eval=FALSE}
 # store mitochondrial percentage in object meta data
-# ctrl <- PercentageFeatureSet(ctrl, pattern = "^MT-", col.name = "percent.mt")
-# ctrl <- SCTransform(ctrl, method = "glmGamPoi", vars.to.regress = "percent.mt", verbose = FALSE)
+ctrl <- PercentageFeatureSet(ctrl, pattern = "^MT-", col.name = "percent.mt")
+ctrl <- SCTransform(ctrl, method = "glmGamPoi", vars.to.regress = "percent.mt", verbose = FALSE)
 ```
 
+
 ## Recalculate Dimensional Reductions
 
 We have now gotten a new Assay added to the Seurat object.
 All we need to do now is to run PCA and UMAP for visualization in lowD.
 We're also going to save the elbow plot for comparison with the log-normed dataset.
 
 ```{r sct_umap_calc}
-DefaultAssay(ctrl) <- "SCT" # explain
+stopifnot(DefaultAssay(ctrl) == "SCT")
 ctrl <- RunPCA(ctrl, verbose = FALSE)
 ctrl_sct <- ElbowPlot(ctrl, ndims = "30")
 ```
@@ -127,7 +133,9 @@ ctrl_ln + ctrl_sct
 What has changed and what hasn't?
 
 > 🧭✨ Poll: [What amount of standard deviation does PC1 explain after sc-transform?](https://PollEv.com/multiple_choice_polls/AQwroH3oNpZ2uQRcly2eO/respond)
-> Hint: if you'd like to determine it numerically, you can use the `Stdev` accessor function to access the standard deviations for PC components in a Seurat object.
+> Hint: Function `Stdev(object)` returns the absolute values of standard deviations for principal components. It's the analogous to `object@reductions$pca@stdev`.
+
+
 
 ## Revisit UMAP
 
@@ -235,7 +243,7 @@ head(b.interferon.response, n = 10)
 ```{r DE_plot,fig.height=4, fig.width=8}
 Idents(immune.combined.sct) <- "seurat_annotations"
 inf_genes <- rownames(b.interferon.response)
-DefaultAssay(immune.combined.sct) <- "SCT"
+DefaultAssay(immune.combined.sct) == "SCT" # integrated
 FeaturePlot(immune.combined.sct,
  features = inf_genes[1], split.by = "stim",
  cols = c("grey", "red")