How do I fetch lincRNAs from Ensembl?

Here’s a very short post on how to fetch lincRNAs from Ensembl using R and the biomaRt package. For those who are not familiar with biomaRt, you can check out my older post on biomaRt. Firstly, start R and install the biomaRt package from Bioconductor by copying and pasting the code below:

source("http://bioconductor.org/biocLite.R")
biocLite("biomaRt")

Now let’s setup biomaRt to use the Ensembl genes catalog for Homo sapiens; the coordinates are based on the latest reference genome:

library("biomaRt")
ensembl <- useMart("ensembl",dataset="hsapiens_gene_ensembl")

The getBM() function is the main query function in biomaRt and consists of three parts: filters, attributes, and values. Using the filter we can obtain genes that belong to the RNA biotype we want, which in this case are the lincRNAs. I will also save various attributes, such as the transcript ID, transcript length, and information about the exons.

lincrna <- getBM(attributes= c('ensembl_gene_id',
                               'ensembl_transcript_id',
                               'transcript_length',
                               'chromosome_name',
                               'transcript_start',
                               'transcript_end',
                               'ensembl_exon_id',
                               'exon_chrom_start',
                               'exon_chrom_end',
                               'strand'),
                 filters = "biotype",
                 value = 'lincRNA',
                 mart = ensembl)

#sort by chromosome, then by start
lincrna <- lincrna[order(lincrna[,'chromosome_name'],lincrna[,'transcript_start']),]

#how many lincRNA transcripts are there?
length(unique(lincrna$ensembl_transcript_id))
#[1] 12211

#I only want lincRNAs on assembled chromosomes
my_chr <- c(1:22,'X','Y')
lincrna_assembled <- subset(lincrna, lincrna$chromosome_name %in% my_chr)
lincrna_assembled$chromosome_name <- factor(lincrna_assembled$chromosome_name,
                                            levels=my_chr)

#how many lincRNAs?
length(unique(lincrna_assembled$ensembl_transcript_id))
#[1] 12157

#create a temp object for plotting purposes
tmp <- lincrna_assembled[,c('ensembl_transcript_id','chromosome_name')]
#remove duplicated rows
tmp <- tmp[!duplicated(tmp),]

#plot the distribution of lincRNAs on the chromosomes
plot(table(tmp$chromosome_name),
     ylab='Number of lincRNA',
     xlab='Chromosome',
     main='Distribution of Ensembl lincRNAs on assembled hg38 chromosomes',
     ylim=c(0,1200))

#the table of lincRNAs will be saved
#inside a file called my_lincrna.tsv
#to find out the location of this file
#type getwd() into the R console
write.table(lincrna_assembled,
            file = 'my_lincrna.tsv',
            quote = F,
            sep = "\t",
            row.names = F)

#save image
save.image(file='lincrna.RData')

For information on the lincRNA biotype in Emsembl, see this page.