qPCR RNA Seq -where to start?

[aoifemc]

Could anyone give me any advice on what proportion of DE genes would need to be validated using qPCR in an RNA Seq study with +8 biological replicates and if its common to just choose those above a certain fold change or would this bias results.
Are there high-throughput ways to do this kind of qPCR for a large group of genes?

Thanks in advance for any help!

[steinmann]

Start by trying to ask the right questions. Then the answer should become obvious.

The answer will depends on what you want to do with these data. qPCR validation may not be needed at all.

[ZWB]

I'm working on a similar problem. We were able to find >200 DE genes in our system and obviously we couldn't do qRT-PCR on all of them. We chose to look at 10 genes (5 up-regulated and 5 down-regulated) to verify our results. I think with 8 replicates you might be able to get away without doing qPCR at all as long as you chose a relatively stringent method to call your DE genes.

[bruce01]

I don't see why qPCR is needed to validate an RNAseq dataset.

A question: why do you do statistical testing to identify significantly up or down-regulated genes, and use false discovery rates, if you then do not believe the results and have to validate them? Why do you want to validate with a method that is as likely, if not more likely, to give spurious results (pipetting errors, comparing against standards, non-quantitative etc). I don't think papers are rejected anymore based on not validating several of your/'the literatures' pet-genes.

The whole idea of high-throughput methods (to my mind) is to get a global view, so why bring it down to a gene-by-gene level again following the HT method?

These are just my own reasons for not doing qPCR validations though

A high-throughput way to validate RNAseq? Microarray? Another RNAseq experiment? Maybe a MiSeq run with target-enriched library for a subset of your data to confirm some of your results at a better coverage?