Hi, I am new to batch effect correction and apologize if this is a naive question.
My goal is to determine how similar my cell line is to the same cell line from an online database. I have the RNA-seq profiles for both samples, but they are obviously from two different batches, and any possible biological differences are confounded with batch.
A colleague of mine suggested that I can still batch effect correct to check similarity, by simply including other cell lines from the online database. So batch 1 would be my sample of cell line A, and batch 2 would include samples of cell line A, B, C, D, and E.
I used CombatSeq for this, and did not tell CombatSeq the identities of the cell lines. After batch effect correction, I see that indeed my sample of cell line A is now closer in the PC plot to the online database's sample of cell line A. Is this analysis valid?
Thanks!
Hi, this is not related to Bioconductor directly. And no, you cannot correct anything because in your own batch you only have a single sample without any representation from other batches. I would also by very! speptical to claim that cells are similar or not based on PCA distance. Cell lines will have all sorts of culture and clonal expansion artifacts, passage numbers, contaminations etcetc that might influence all that.
Thank you! And apologies for posting here. Do you know which forum I should go to in the future for questions like these?
https://www.bioconductor.org/packages/release/BiocViews.html#___BiologicalQuestion