Inefficient Manual Coding Required

[DarioS]

Some of the code in the vignette is inefficient and perhaps should be convenience functions for users. For example,

gene.set.list <- list()
for (gene.set.name in unique(mdb_c5_bp$gs_name)) {
    gene.set.list[[gene.set.name]] <- mdb_c5_bp[mdb_c5_bp$gs_name %in%
        gene.set.name, ]$gene_symbol
}

is inefficient and not how R lists should be created. A better way is split(geneSets[["gene_symbol"]], geneSets[["gs_name"]]).

Also, Preparation section has NormalizeData(pbmc3k) but doesn't explain why it is required, What is the purpose? It seems to be repeated again in MCA step, which seems inefficient to me.

> embeds <- compute.mca(allHumanSeurat)
Normalizing layer: counts
Performing log-normalization

Why does log-normalisation need to be done twice?