Sequencing Single-end on DE to save money, good idea?

[Kate.W]

Hi people!

We assembled a de novo assembly using deep PE reads. Now we want to run biological replicates to analyse DE. Naively, I thought we should run he very same PE sequencing on those replicates, but I was advised by the people using illumina that, as my transcriptome is assembled and seems to be of good quality, I could go for SE sequencing. It'd save me a lot of money for same results. What would you think about that? Thanks!

Kate

[swbarnes2]

They are probably right. If you aren't looking for alternate splicing, or SNPs, two paired reads on a transcript won't tell you more than one read.

[alexdobin]

It does depend strongly of the biological questions you are trying to answer. If you want to analyze just the differential gene expression, the SE would be almost as good as PE - the main difference is that PE reads will contain fewer multi-mappers and will allow you to "resolve" better highly paralogous genes.

On the other hand, if you want to quantify alternative isoforms, or assemble novel isoforms, PE reads give you much more power than single end. Many isoform quantification algorithms rely on "insert size distribution" which you can only get from PE reads.

All in all, I think PE reads are more versatile than SE.