How do you select DEGs for further validation?

Question

[marina.wakid]

Perhaps this a silly question, but I'm overwhelmed with the different ways to give context to results after DGE analysis (RNA sequencing in a disease vs healthy control context). Of course I know that one should establish criteria for significance and/or log fold change, but what does one do after? Does anybody have an approach they swear by?

To clarify: I would like to select a handful of DEGs for further experimental validation using RNAscope and qPCR but ideally the selected DEGs would be "cohesive" with one another. Do people go for genes returned by interesting terms after ORA, or do people prefer to select genes whose proteins interact (informed via STRING analysis)? Or do people simply prefer to let previous literature inform their selection? Or do people select hub genes from WGCNA modules?

Thank you!

Marina

Answer 1

[Mensur Dlakic]

Your problem is very common among those doing DEA. Most people are not prepared for hundreds of candidates that sometimes emerge from those studies.

All the approaches you outlined above have merit. It depends on what one is most interested in studying. If you are more interested in confirming known relationships, the literature will help you. If you want novel relationships, they presumably are not in the literature. I would always look into STRING clusters because it is easy to do. But really, checking out all of the above options is easier than doing the actual experiments. There is nothing stopping you from doing all computational analyses and literature search, and integrating them into a coherent hypothesis before you move on to additional experiments.

[marina.wakid]

Thank you so much for your feedback! I feel more reassured know :)