get_human_orthologs.R

get_human_orthologs.R

library(biomaRt)
species_tx2gene_map = c(
    "btaurus",
    "mauratus",
    "mmulatta",
    "ggallus",
    "sscrofa",
    "mmusculus",
    "dreiro")
    "rnorvegicus",
    "pabelii",
    "mdomestica",
    "ptroglodytes",
    "mmusculus",
    "mpfuro",
    "oanatinus",
    "ggorilla",
    "ppaniscus")

attributes_base <- c("_homolog_ensembl_gene",
                     "_homolog_associated_gene_name",
                     "_homolog_ensembl_peptide",
                     "_homolog_chromosome",
                     "_homolog_chrom_start",
                     "_homolog_chrom_end",
                     "_homolog_canonical_transcript_protein",
                     "_homolog_subtype",
                     "_homolog_orthology_type",
                     "_homolog_perc_id",
                     "_homolog_perc_id_r1",
                     "_homolog_goc_score",
                     "_homolog_wga_coverage",
                     "_homolog_dn",
                     "_homolog_ds",
                     "_homolog_orthology_confidence")

mart <- useMart("ensembl", dataset = "hsapiens_gene_ensembl")
z <- listAttributes(mart)
species <- species_tx2gene_map[1]
for (species in species_tx2gene_map) {
    attributes <- c('ensembl_gene_id', 'external_gene_name', paste0(species, attributes_base))
    attributes <- intersect(attributes, z$name)
    filter <- paste('with', species, 'homolog', sep = '_')
    orth <- getBM( attributes,filters=filter,
                   values=TRUE, mart = mart, bmHeader=FALSE)
    df <- as.data.frame(orth)
    colnames(df) <- gsub(pattern = paste0(species, "_"), replacement = "", x = colnames(df))
    write.table(df, file=file.path('data/orthologs/', paste0('human_', species, '.tsv')), row.names=F, col.names=T, quote=F, sep='\t' )

}