Include other covariates

[plbngl]

Hi Natsuhiko,

For my caQTL analysis, in addition to the covariate file generated from PCA of the peak matrix, I would like to add other co-variates, such as male/female and PCA from genotype data (ancestry). How shall I do that? Shall I combine them in a unique file or do they have to be separate files?

Another question - in some cases I have outliers in the PCA[1 and 2] of the peak matrix - but I do not want to remove the samples because I only have 10 samples in total - how would you recommend to proceed in these cases?

Thank you very much for your great work!!

Paola

[natsuhiko]

Hi Paola,

You can combine your covariates (such as Sex, genotype PCs and so on) with the PCs calculated from the feature count matrix. Please see "R/makeCovariates.R" and modify it to generate your own covariates file. You can keep the outlier in the data, because covariates adjust the feature counts. However, those outliners may not contribute the power to detect QTLs, anyway.

Best regards,
Natsuhiko

[plbngl]

Thank you for your answer,

If I understood correctly the makeCovariates.R function returns selected PCs from fpkm count matrix so I am not sure why I need to modify the function and not just append my other PCs from genotype and sex to the output?
I might not have understood then...

Thank you again for your help!!
Paola