I am trying to learn how to analyse normalised DIA-MS data and I am struggling with it ://

The original dataset I got is (6 conditions (2 samples each)) with 3 technical replicates (total: 36 sample columns x 1128 proteins in rows). I removed the proteins with more than 80% of NA, and impute with kNN the rest of missing values.

Then I collapse the replicates doing the average of the 3 replicates (is this correct?)

So now, I have a dataframe with 12 columns x 1072 rows.

The question is... I am not sure how to perform the differential analysis from here (assuming I did well the previous steps lol)

Could someone help me with the tools or code to compare all the 6 conditions between them?

I read something about Limma package, but not sure if it can be used in this DIA-ms data and neither sure of how to use limma for it (I already check the vignette with no succeed)

UPDATE: I tried to use Limma as follows:

fit1 <- lmFit(norm.data_collapsed, design)
fit2 <- contrasts.fit(fit1, contrasts = contrast)
fit3 <- eBayes(fit2)
limma.res.01 <- topTable(fit3, coef = comb.contrast[1], n= Inf)

This returns me a non-sense values for logFC or significance:

head of the results topTable

Any help would be very welcome,

Best regards


1 Answer 1


Regarding the big FC values, I just realise the Limma lmfit function requires as input a log2 matrix... so I just transformed my data and now makes more sense. Still I have the doubt regarding the collapsing technical replicates and and if some one can verify the pipeline would be like that.



  • $\begingroup$ Please could you mark this answer as accepted? That would be very welcome. $\endgroup$
    – M__
    Commented Mar 15 at 19:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.