I have performed gene expression analysis using transcriptomic data from RNA-seq technology. I have validated some of my results for some genes by using qPCR and western blot. Although I am comfortable with stating that western blot is an orthogonal validation method for RNA-seq data, I am not so certain how to define validation by qPCR. Is it parallel or orthogonal validation? I would think since RNA-seq and qPCR are similar (use of RNA converted to cDNA) that performing the latter as validation for the former would be a parallel validation. However, these two methods are certainly different enough to suggest that perhaps qPCR is orthogonal after all. The literature seems to be mixed, but perhaps the answer is in the details. Some clarity on this is appreciated.
I would say qPCR is a parallel validation in gene expression analysis. It is not at all independent like a Western blot to confirm the biological result of a study. That too, in RNAseq, whether qPCR validation is required for confirming fold changes is debatable as it has been proved by many publications that there exists a tight correlation between the 2 methods. In gene expression analysis, qPCR is just a technical validation unless you are using it in other biological samples that you do not want to do RNAseq for, to reduce cost.
But it is indeed useful and can be considered orthogonal in case of re-affirming novel transcripts and alternative transcripts to rule out false positives that can crop up in the insilico analysis (identifying alternative transcripts is quite difficult and even the best tools give errors). In this case, qPCR gives an independent and biological validation of your insilico identified alternative splicing event.
I would say qPCR is a parallel validation in gene expression analysis. It is not at all independent like a Western blot to confirm the biological result of a study. That too, in RNAseq, whether qPCR validation is required for confirming fold changes is debatable as it has been proved by many publications that there exists a tight correlation between the 2 methods. In gene expression analysis, qPCR is just a technical validation unless you are using it in other biological samples that you do not want to do RNAseq for, to reduce cost.
But it is indeed useful and can be considered orthogonal in case of re-affirming novel transcripts and alternative transcripts to rule out false positives that can crop up in the insilico analysis (identifying alternative transcripts is quite difficult and even the best tools give errors). In this case, qPCR gives an independent and biological validation of your insilico identified alternative splicing event.
Thank you for your response Abhinay. It would seem that my situation is part parallel and part orthogonal for qPCR since I used other biological replicates for my ex vivo samples (orthogonal), but compared with the same cell line samples. Perhaps if I had used a separate set of replicates for my cell line samples my qPCR validation would be fully orthogonal. Though there is a debate about how you would practically get true replicates of a cell line that has been derived from a single cell clone, but that is a discussion for another day. Thanks again.
There are some rules, unfortunately not too widely known, to make the qPCR technically more comparable to the RNAseq fold changes. Eg:
* the design of primers should pick a region which has the stable coverage of reads on RNAseq (can be checked with IGV).
* The primers should be placed in the same exon, which also does not have there shorter exonic forms in known databases (Ensembl, AceView)...
Otherwise qPCR has "artefacts by design". Generally - I trust more RNA-seq. it is more general and less "artefact prone".
- Badges
- Science method