Load RNA counts object, that is a genes by cells matrix with gene names on rows and cell barcodes on columns.


In roder to create the CONGAS+ object, we need to pool together counts across each genomic segment. Thus, we need to have a dataframe that maps each gene to the corresponding coordinates. In some specific settings, such as 10x multiome sequencing data, this information is already available. However, when needed users can exploit Biomart to map each gene to the corresponding genomic position, as we show below. This will produce the variable that is needed by the function to generate the input data in the format required by CONGAS+.

ensembl <- useEnsembl(biomart = "genes", dataset = "hsapiens_gene_ensembl")
z <- getBM(c("hgnc_symbol", 'chromosome_name','start_position','end_position'), 
             filters = "hgnc_symbol", 
             ensembl) %>% 
             dplyr::rename(gene = hgnc_symbol, 
                          chr = chromosome_name, 
                          from = start_position, 
                          to = end_position) %>%
              filter(chr %in% c(seq(1:22), 'X', 'Y')) %>%
              mutate(chr = paste0('chr', chr)) 

z = z[!duplicated(z$gene),]

gene_names = rownames(rna_counts)

features = tibble(gene = gene_names) %>% left_join(z) 
#> Joining with `by = join_by(gene)`