significant differential expressed genes from two microarray accession

Question

[zizigolu]

hi,

in GSE14407 datasets there are normal and cancerous samples but in GSE9891 there are only cancerous samples but both from the same platform. is it wise to get normal samples from GSE14407 combined with GSE9891 because i am going to extract differential expressed genes between cancer and normal.

may someone share comments please

thank you

[Mike]

I think, Yes, you can do that.

First, download all raw CEL files for GSE14407 & GSE9891, copy into a single folder, and do normalization (rma..etc) for all samples together, and then go for DE genes.

[devikaparvathy]

Is this protocol still valid? Can i use control and test samples from different experiments but same platform to run a DE analysis? Do I have to use ComBat or sva to adjust for variations???

Also is there any reference paper that uses this methodology foolproof?

Thank you in advance.

[WouterDeCoster]

Can i use control and test samples from different experiments but same platform to run a DE analysis?

No, you cannot. You cannot differentiate the technical variability from the biological variability. This is not a valid analysis.

[devikaparvathy]

What if I I am using that data for a following meta-analysis?

Saw the following paper that uses such an approach - https://academic.oup.com/nar/article/45/17/9860/4084660 - see Table 1/SCNvsWB)