"Bias" in NGS sequencing processing

Question

[Andrew Liu]

Hi, I'm now considering a confusing question for next-generation sequencing. Suppose I have a cell with 4 genes, and with a consistent expression quantity of 1:1:1:1. In wild type, their expressions are 50:50:50:50 and in mutant are 500:500:500:500. But because of the sequencing method, samples need to be extracted and PCR, then select some for sequencing mechine. And finally I get some DNA to sequence. Here comes a question, if I get 1G data from wild and mutant sample, but they may show the sample expression quantity of 5:5:5:5. How can I cope with the problem like this or in the realitistic data? Thank you!

Answer 1

[igor]

With NGS, you probably would not be measuring only 4 genes. It's not very common, but it's possible to perform targeted RNA-seq for specific genes of interest. In that case, you would also have housekeeping genes to use as controls. That would be analogous to more traditional methods like qPCR.

Answer 2

[swbarnes2]

Well, yes...

Crappy experimental design -> meaningless data -> bad conclusions.

So, don't have crappy experimental design. If you really only care about 4 genes, do qPCR with a control gene. Or, if you want a targeted panel, make sure it is big and diverse so you have some kind of background you can compare to.