How Cufflinks Handle Same Gene Very Differently Expressed Between Two Biological Replicates[Rna-Seq]

Question

[Sukhi Singh]

Hi guys,

How cufflinks handles the genes whose expression values varies a lot among biological replicates. If it takes into account, is it averaged then or what? Also, imagine a case when the fold change between biological replicates is bigger than the conditions to compare?

Another question is, why the same gene in the same condition is having slightly different RPKM values in different lists? Consider, comparing condition A and B, gene X has RPKM=23 and condition A and C, same gene X of A has RPKM=22.5.

Thanks

P.S. I think I need to study more about cufflinks.

Answer 1

[Istvan Albert]

This question is one that has bee hounding bioinformaticians since the beginning of times. The right normalization and data scaling are essential to producing reproducible results. Here is one paper Evaluation of statistical methods for normalization and differential expression in mRNA-Seq experiments (BMC Bioinformatics)

That being said if one had only two biological replicates their options are very limited when it comes to figuring out what the right value could be.

[Sukhi Singh]

Agree Istvan, and thanks for the paper, but do you some comments on how cufflinks handle this.

[Istvan Albert]

See this page

http://cufflinks.cbcb.umd.edu/howitworks.html#hdif

[Sukhi Singh]

Thanks Istvan :)