why appears transcritps with 0 expression if it was used for the novo transcriptome assembly?

Question

[Pilar]

I've done a de novo transcriptome assembly using Trinity pipeline. I've used abundance_and_estimation.pl script with Kallisto for quantifying. My question is, why transcripts with 0 counts among all replicates appear? If they were used for the assembly, they all should be quantified.

Thank you in advance, :)

[GenoMax]

it was used for the novo transcriptome assembly

Transcript was not used for the assembly, rather it was a product of that assembly. It is possible that there are alternate isoforms of this particular transcripts that are favored by kallisto when assigning reads. Have you checked that? Have you also made sure that your assembly is non-redundant?

[Pilar]

You're right. When I'm talking about transcripts at first, I was meaning reads. In the matrix of counts obtained from Kallisto, there are 5826 isoforms with 0 expression value (and 31368 that have expression in 8 replicates).

• 0 -> 5826
  • 1 -> 7992
  • 2 -> 11273
  • 3 -> 12815
  • 4 -> 13713
  • 5 -> 13624
  • 6 -> 13526
  • 7 -> 13708
  • 8 -> 31368

Answer 1

[lieven.sterck]

[ Not sure if this is THE answer but it's how I see it at least. ]

Well, you have to take into account that it is the result of an assembly effort. Reads are grouped on overlap/similarity and from the 'group' a consensus sequence is made (consider it a, potentially large, MSA that is being collapsed into 1 sequence). Taken this into account and combining with the fact that in most organisms you will have heterozygosity and allelic variation it is possible that you end up with an assembled sequence (consensus) that does not exists in real life, meaning that it can have a combination of allellic and other variations that is not 'real' ... as a result it might be that due to this none of the reads map to it anymore as none of them have the exact same set of variations in a single read.