I want to ask if short RNA fragments caused by degradation during RNA extraction/processing can influence the quantification process and if the genes would get more counts/reads aligned due to the presence of these fragments.
If so, is there a method to account for that? taking into consideration that some DEA packages like DESeq2 require the raw non-normalized counts to perform the analysis.
If your question is whether poor quality RNA can negatively impact {everything}, then yes. The s in sout principle applies to RNA-seq as well. Check by PCA if there are any signs of batch effect that can be corrected for. Or maybe use somethign like RUVseq or SVA to estimate factors / surrogate variables to capture unwanted variation.
Many thanks for your answer, actually the PCA plot looks fine. I was worried that having shorter fragments of a gene could give higher counts of that gene during quantification, causing false-positive upregulation of that gene.
RNA degradation can indeed have quite a large effect - and the problem mainly is that it is a systematic error where long transcripts are more affected than short ones. In the end that leads to various kinds of imbalances.
One thing that you could potentially do is to count over a shorter and fixed distance rather than the full length of the transcript. Typically, the interval would be around the end of the transcripts because the Illumina strand-specific library prep has a reverse complement step.
What this means is that you could count over the last 500bp of the transcript. Alas many additional issues arise because now you need to account for exons etc. I know this because we tried the technique once and it was surprisingly more complicated to get it right than anticipated.
Your mileage may vary. In the end what matter is just how pronounced is the effect. How many transcripts are affected.
If your question is whether poor quality RNA can negatively impact {everything}, then yes. The s in sout principle applies to RNA-seq as well. Check by PCA if there are any signs of batch effect that can be corrected for. Or maybe use somethign like RUVseq or SVA to estimate factors / surrogate variables to capture unwanted variation.
Many thanks for your answer, actually the PCA plot looks fine. I was worried that having shorter fragments of a gene could give higher counts of that gene during quantification, causing false-positive upregulation of that gene.