Jochen, i'm well aware of the pitfalls of RT-qPCR, in fact most of my posts have been about this.

In terms of experimental procedure and data analysis, there are less steps for introduction in errors for RT-qPCR when compared to NGS and to some aspects microarray. There are excellent, widely established guidelines for wide ranges of RT-qPCR topics starting from cDNA synthesis, primer design and data normalization. You should know this because you have written about this yourself to some extent.

As NGS is a constantly evolving field, such topics do not exist for NGS. I have watched a colleague of mine struggle during his PhD with the NGS platform he had (Proton sequencer) with many technical aspects of the topic including the machine itself. 

When i say qPCR is easy to use, i meant it is easier to use than the high-throughput counterparts. You must remember that you are giving advice to layman here, not someone who has been for 12 years working in the field. It is possible that their supervisors aren't even molecular biologists. Regardless, if you plan to present your data to a journal, odds are you will likely need some confirmation. We've gone as far to validate our microarrays results with RT-qPCR and Western blotting. This kind of validation shows that you understand the pitfalls of your platforms and techniques, not lack of confidence in it.

No matter how you customize RNA seq platforms, the RNA-seq only provides a “broad scope” of the genes at hand. This is excellent to examine the effect of a large number of genes (high coverage), detect alternative splicing, small RNA characterizations etc. But the data obtained is broad and relative and also the sensitivity is based on the read depth (50M reads, 100 M reads etc) or the length of the targeted genome. The latter is important because when using longer/larger reads the detection sensitivity may decrease.

Qpcr is limited to studying few genes at a time, however it can be utilized to quantify a set of genes that might be sensitive. Thus we can understand gene quantification at a more specific level.

If your starting knowledge is “I have no clue which gene is relevant from the whole genome” … then RNA seq will help to narrow it. But it will narrow it down to few pathways and few different genes. Thus in order to progress your study more specifically, the more specific qpcr method can be used. This will now help you to narrow down your research interest and help to build a better story line.

Due to errors that might be generated during RNA seq platforms, reviewers now do require confirmation by qpcr. Narrow down your interest and perform only those genes.

I personally think that confirmation of expression data coming from high throughput techniques (like Microarray and RNAseq) by means of RT-PCR is not strictly mandatory, but it is however highly recommened. If primers are carefully designed, amplifying only specific gene, if the amplification efficiency is high, RT-PCR should provide the most accurate abundance of a particular target in a gene expression study.

In the past, expecially for Microarrays, the higher dynamic range of RT-PCR was necessary to verify DEGs.

Regarding RNA-seq, the dynamic range is dependent on the depth of sequencing while microarray has more or less a fixed dynamic range. If your data set came from a good run with sufficient depth I guess there would be no significant difference in medium to high expression genes between RNA-seq and RT-PCR, however, I think that the DEGs with the lowest and the highest expression profiles should be checked by PCR..

Good luck,

Alessandro

As the previous two people have explained, it is not necessary but deemed good practice.

High-throughput methods such as microarrays and NGS present inherent difficulties in data analysis which must be carefully analysed by an expert. This has given rise to the need of an bioinformatician.  For example, microarrays require "significant analysis of microarray" whilst NGS require assembly, which require some expertise. Remember that these technologies were/are at the forefront of technology and there aren't many well-established protocols.

RT-qPCR employs a different chemistry to the high-throughput machinery and is widely established. easy to use, and interpret. 

I don't agree to Can that qPCR is "easy to use and to interpret". Too many people fail to correctly analyze and interpret qPCR data. Too many people torture their qPCR data long enough to just show the effects they want them to show (at least among those I know!). Reference genes are badly established, invalid amplification curves are used to get ct values, reactions are not checked for the co-aplification of unspecific products (especially when sequence-specific detection probes are employed), quantifiactions are done on unstable reactions or on reactions with apparent amplification efficiencies that are considerably lower or even higher than 100%, and so on. And on top there are many papers showing non-sensical calculations (if they are shown at all!), like analyzing the ratio of ct values and alike.

I don't see much value in a technical confirmation of the results. If you use high-throughput methods to get the "big picture", to find regulatory networks and pathways, the evidence from the high-throughput data is always far better than the additional information you get from confirming some few candidates. If you are interested in some particular candidates (as possible drug targets or whatever) then a functional assay would be much more interesting and insightful. If such a functional assay is much more expensive and time-consuming than the establishment of valid qPCR assay (including the search for reference genes and their validation!) then a qPCR may help to save money and time as one will reduce the (expected) number of functional assays performed on "false-positive" candidates.

Jochen, i'm well aware of the pitfalls of RT-qPCR, in fact most of my posts have been about this.

In terms of experimental procedure and data analysis, there are less steps for introduction in errors for RT-qPCR when compared to NGS and to some aspects microarray. There are excellent, widely established guidelines for wide ranges of RT-qPCR topics starting from cDNA synthesis, primer design and data normalization. You should know this because you have written about this yourself to some extent.

As NGS is a constantly evolving field, such topics do not exist for NGS. I have watched a colleague of mine struggle during his PhD with the NGS platform he had (Proton sequencer) with many technical aspects of the topic including the machine itself. 

When i say qPCR is easy to use, i meant it is easier to use than the high-throughput counterparts. You must remember that you are giving advice to layman here, not someone who has been for 12 years working in the field. It is possible that their supervisors aren't even molecular biologists. Regardless, if you plan to present your data to a journal, odds are you will likely need some confirmation. We've gone as far to validate our microarrays results with RT-qPCR and Western blotting. This kind of validation shows that you understand the pitfalls of your platforms and techniques, not lack of confidence in it.

Dear Parviz,

"We have RNA-seq data set and we selected some genes involved in hormone pathways for verification the RNA seq results but some genes are not in DEGs list."

How close these genes are in the RNA seq statistical rank or expression pattern/values compared to the ones in DEGs list confirmed by qPCR? You can list them together in a table with the raw p values and provide it in supplementary information. Let the readers/reviewers see the whole analysis and judge. 

This is a question I hear a lot. I usually refer people to this paper (http://www.jbiol.com/content/8/1/3). The important thing to remember is that RNA-seq is used to get a global picture and that's that. Get an idea of the general processes that are changing. If the whole paper following the RNA-seq will be entirely based on the finding that a single gene changed, as observed via RNA-seq, then sure, it may be worth taking a look with new samples and a different technique. In general, as a reviewer, I don't ask for such 'validation' data. As Jochen mentioned, functional data, consistent with the gene expression data obtained from the RNA-seq data, is way more convincing and useful.