How to calculate gene expression level (RPKM) in RNA-seq with BED input file

[luuloi]

I have a read aligned file with format as BED file (of the whole genome):

assembly strand start end sequenceA sequenceB id
Y + 2718307 2718348 TGCGTGGGGCCTTAACTTGGCCCAGCCTTGCGGCAGCGATTC 1
Y + 2905228 2905269 CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGGGGTTTTTTTT 2
Y + 2905228 2905269 CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGGGGTTTTTTTT 3
Y + 2905228 2905269 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAGAGGGTTTTTTTT 4

I want to calculate the expression level in RPKM. Could anyone can help me, please?
Thanks,
Loi

[Thias]

RPK = Reads per kilo base (of transcript)
RPKM = RPK per million (total reads)

Example calculation:

140 mapped reads (28000000 mapped reads total), annotated transcript 7000bp
RPK = 140/7 = 20
RPKM = 20/28 = 0.71

------------------------

This example shows you, that the information provided is not sufficient. First you need to know to which genome assembly your data is mapped to. Then you need to pick the transcript of your choice (gene of interest) and retrieve the genomic coordinates for it's transcript. (only exons, some additionally neglect the 5'UTR). Subset your BED file to retain only the reads which overlap your annotated transcript. The rest of the calculation is done like shown above.

If you are interested in not only in one particular RPKM but in a genome wide result, then you should check what your bioinformatic language of choice offers regarding packages and libraries. Whether you use R, Python, Perl, Go or some other, there should be stuff around for any of those.

BTW: BED is a rather unusual format for mapped reads and usually more commonly used for graphical tracks to be displayed e.g. in the UCSC Genome Browser or IGV. Maybe it would be worth a try to find out which mappinng tool they used for the alignment and consult the manual of this software how to proceed with the output.

[luuloi]

Thank you Thias for your useful information!

I interested in for the whole genome and mainly use Python, however I can not find any available package/script doing this job!

In their publication, they only mentioned that: "RNA-Seq read sequences produced by the Illumina analysis pipeline were aligned with the TopHat software and reference genome is hg18" and gave these files (in this BED like format) without further information.

And what I need also the expression of each gene, so there will be a python script or R script elsewhere can calculate RPKM with input file as BED like format?

[Thias]

If you use Python why not give rpkmforgenes.py a try ? Didn't try it out myself, but it seems they even want .bed input exactly in the way you have it.