NewAEMSet requires the genes matrix to be in order 1: ensembl_id then 2: gene_name

[quanaibn]

if changing the order, there is a error: Error in $<-.data.frame(*tmp*, "control", value = TRUE) : replacement has 1 row, data has 0 I think this can be solved by calling columns by names?

[asenabouth]

Could you please post a snippet to show how you came across the error? Thanks

[quanaibn]

exprs_mat <- read.csv(paste0(path, 'mmThymus_scRNA_Aggr_ExpressionMatrix_V4.csv'), header=T)

the exprs_mat has 3 batches, with suffix .1 to .3; and has a gene_id column with gene symbol and another row.names column with Ensembl ID

exprs_matrix <- exprs_mat[,3:ncol(exprs_mat)] barcodes <- colnames(exprs_matrix) batch.information <- lapply(strsplit(barcodes, '\.'), [, 2) barcodes <-as.data.frame(barcodes) colnames(barcodes) <- c("cell_barcode") barcodes$batch <- as.numeric(batch.information)

genes <- exprs_mat[,1:2] colnames(genes) <- c("ensembl_id", "gene_name") gene.names <- make.unique(as.vector(genes$gene_name)) rownames(exprs_matrix) <- gene.names genes$gene_name <- gene.names

mito.genes <- rownames(exprs_matrix)[grep("^Mt-", rownames(exprs_matrix), ignore.case = TRUE)] ribo.genes <- rownames(exprs_matrix)[grep("^Rps|^Rpl", rownames(exprs_matrix), ignore.case = TRUE)] control.list <- list(Mt = mito.genes, Rb = ribo.genes)

expression.matrix <- exprs_matrix

expression.matrix <-exprs_matrix gene.information <- genes[,c(2,1)] cell.information <- barcodes

aem.set <- NewAEMSet(ExpressionMatrix = expression.matrix, CellInformation = cell.information, GeneInformation = gene.information, Controls = control.list)

The NewAEMSet error comes from here: gene.information <- genes[,c(1,2)]
If i swap it  gene.information <- genes[,c(2,1)], the NewAEMSet run well

[asenabouth]

Please use the inbuilt function ConvertGeneAnnotation to change the annotations. This will update the control names in the object, move the columns around and sync all the slots together.