Running PEER regression with MOFA2

[MarcElosua]

Hi @rargelaguet

Thank you so much for developing MOFA2, such an amazing package!

I was trying to use PEER but saw this issue where you commented on how MOFA with 1 view was basically PEER.

I'm trying to re-run a previous analysis that had been run with PEER to regress out the effects of covariates of interest from gene expression data as shown below:

##################
#### Run PEER ####
##################
K <- 10
model = PEER()
PEER_setPhenoMean(model, as.matrix(t(norm_expression)))
PEER_setAdd_mean(model, TRUE)
PEER_setNk(model, K)
PEER_getNk(model)
PEER_setCovariates(model, as.matrix(covariates))
PEER_setNmax_iterations(model, 10000)
PEER_update(model)

# Save residuals
residuals = PEER_getResiduals(model)
colnames(residuals)  <- row.names(norm_expression)
row.names(residuals) <- colnames(norm_expression)
dump <- data.frame(ids = row.names(norm_expression), t(residuals), check.names=FALSE)

Is there a straight forward way to reproduce this with MOFA2?

Thank you very much for your help!