Can we use exon- counts to calculate coverage?

Question

[unawaz]

Hi all,

I was wondering if you could help me with this. I currently don't have access to the bam or fastq files of the samples I'm working on from a public database. But I do have exon counts for these. I'm wondering if I can use these to calculate coverage where I can take the average of the RPKM X Length of transcript, which would reflect the total read coverage?

I read it in a paper and was wondering if someone has tried it? I can't seem to find their original code for how they did it.

Thanks!

[Shred]

Do you have raw counts or RPKM ? You need the coverage over exons or transcripts?

Answer 1

[Istvan Albert]

I would say that you can't quite get accurate RPKMs from the exon counts because when summarizing over exons, a read may contribute with counts for different exons.

If you wanted a reasonable estimate, though, then summing up exons might be an acceptable estimate for a transcript count, in which case you can sum the exons for a transcript/gene, and use the resulting as the number of mapped reads (perhaps find a correction factor to reduce this number a bit)

[unawaz]

Thank you! I'm basically looking to quantify the 3′ bias of each gene for each sample but don't have the raw data or aligned data at hand, only RPKM values. Essentially I want to use the counts to obtain a read coverage profile across all exonic positions of a representative reference transcript. How can I do this in R?