Entering edit mode
3.2 years ago
alonziso
▴
20
Hi, I'm not sure if this is the right place to ask but... I'm working with RNA-Seq data of tetraploid wheat and I'm interested in comparing differences in expression between my DEG (calculated using DESeq) to their homeologs. As far as I can understand I cannot use DESeq counts for this because they are not suitable for within-sample comparison. I am not sure if I should use TPM instead and if I do use TPM how can I say something meaningful on the overall differences between treatments (parents and hybrids in my case)?. Thanks.
Hi alonziso,
ok, not an easy case indeed. I'm a "plant-guy" myself so I am somewhat aware of this situation but could still do with a bit more info.
Is it correct to assume it's an alloploid species (so combination of two different genomes) and this not an auto-ploid (same genome doubled)?
and perhaps not the reply you are hoping for but did you look around in literature already? I'm convinced people have looked into this before. The wheat genome paper might be a good starting point already.
Hi, First thing first, thanks for answering... It is an allopolyploid species, although, the hybridization performed in my experiment did not cause the allopolyploidization (I work on emmer so it is a tetraploid species, to begin with). I looked at some papers that did something similar but they had a much larger data set from different tissues, they did an average per tissue for every gene. as I am working on a small data-set with samples from only one tissue it should have been simpler... but I'm not sure how should I approach this, should I preformed a geometric mean across all samples. I am not sure if this is the right way.
Perhaps you can look in the literature for things related to isoform quantification? Not exactly the same thing (far from perhaps) but might still give you some useful insights?
Something like pseudo-quantification with Salmon might work here?
I did use Salmon, so I have the TPM data. but the question remains. is it the right way to perform a geometric mean for genes and then compare them to one another