Convert cufflinks gene ids to ensembl ids

id_conversion.R

#! /usr/bin/env Rscript

suppressPackageStartupMessages(expr = library(package = "rtracklayer"))

reference_granges <- import(con ='file1.gtf')
reference_frame <- unique(x = data.frame(
  ensembl_gene_id = reference_granges$gene_id,
  ensembl_transcript_id = reference_granges$transcript_id,
  stringsAsFactors = FALSE))
rm(reference_granges)

assembled_granges <- import(con ='merge.gtf')
assembled_frame <- unique(x = data.frame(
  gene_id = assembled_granges$gene_id,
  transcript_id = assembled_granges$transcript_id,
  gene_name = assembled_granges$gene_name,
  ensembl_transcript_id = assembled_granges$nearest_ref,
  stringsAsFactors = FALSE))
rm(assembled_granges)

merged_frame <- merge(x = reference_frame, y = assembled_frame, by =
                        "ensembl_transcript_id")

merged_frame[merged_frame$gene_id == 'XLOC_000005', ]

###
# for each gene_id ...for 1000 lines approx 5-10sec
# apply would make it faster...

uniq_gene_id <- unique(merged_frame$gene_id)
ensg_ids <- c()
enst_ids <- c()
for (i in 1:length(uniq_gene_id)){
  xloc_num <- merged_frame[merged_frame$gene_id == uniq_gene_id[i], ]
  ensg_ids[i] <- paste(unique(xloc_num$ensembl_gene_id), collapse=',')
  enst_ids[i] <- paste(unique(xloc_num$ensembl_transcript_id), collapse=',')
  print(i)
}

fill_dtaframe <- data.frame(
  uniq_gene_id=uniq_gene_id,
  ensembl_gene_id=ensg_ids,
  ensembl_transcript_id=enst_ids)

#dta to match
dta <- read.table(file="table_for_conversion.tsv",
                  sep="\t", header=TRUE)
merged_dta <- merge(x=dta, 
                    y=fill_dtaframe, 
                    by.x="gene_id", 
                    by.y="uniq_gene_id")
write.table(merged_dta,
            file="converted_table.tsv",
            sv", quote=FALSE,
            sep="\t")