Hello, Currently working on RNA-seq data, I have total 5 sample, and 4 sample from the unstranded library preparation method my objective to find the differential gene expression is the unstranded data effect differential expression because we will have the count for untranscribed antisense strand in some genes , can any one help how I can resolve this problem i went through many threads from bio-stars but i didn't get any clear ans
a small example how my data is i mentioned below
sample library_prep condition
sample1 stranded control
sample2 unstranded control
sample3 unstranded treated
sample4 unstranded treated
sample5 unstranded treated
In this type situation how can I proceed with my experiment for downstream need some expert suggestion
Thank you
•
written
by
luckysardar171 •
So your question is what to do with the samples prepared with different library prep methods ? I think the easiest answer is to drop the stranded sample. The reason for this is that the extra sample will probably introduce more noise into the analysis than it is worth, due to the library prep effect . To check that you can make a PCA plot of your data and see how much variance is explained by the condition vs library prep factors. The alternative approach is to include the stranded sample in the analysis and control for the library prep effect in DESeq2 design.
Another thing to consider is that by default, most read counter (featureCounts, HTseqCounts, etc) discard reads that can not be unambiguously attributed. Therefore your ability to measure difference of expression for sense-antisense transcripts pairs will be very low., which make sense since your data is mostly unstranded.
Thank you for the reply
I did PCA its looking like (link)
•
written
by
luckysardar171 •
Sorry for the mistake, provided the link for the PCA plot
Log in to answer this question.
