Use of netConstruct with CLR-transformed data

[albaperezcuervo]

Dear NetCoMi team,

I am a regular user of the NetCoMi package and typically use it with raw count data as shown below:

PD.net <- netConstruct(data = physeq.PD,
dataType = "counts",
measure = "spieceasi",
sparsMethod = "none",
nboot = 1000)

PD.props <- netAnalyze(PD.net,
centrLCC = TRUE,
clustMethod = "cluster_fast_greedy",
hubPar = c("degree", "betweenness"),
hubQuant = 0.9,
weightDeg = FALSE, normDeg = FALSE)

plot(PD.props, ...)

However, in my current project, I am working with microbiome data that has been normalized to relative abundances, and then transformed using CLR (Centered Log-Ratio) due to the need to apply ComBat for batch effect correction — which requires CLR-transformed input.

Given that spieceasi within NetCoMi applies CLR internally when using measure = "spieceasi" and dataType = "counts", I am concerned that applying netConstruct() in this case would result in a second CLR transformation, distorting the data.

Is there any way to indicate to netConstruct() (or to the SpiecEasi backend) that the input data is already CLR-transformed, so that no additional transformation is applied?
Or, alternatively, is there a recommended workflow within NetCoMi for working with already CLR-transformed data, especially when ComBat correction has been applied?

I appreciate any guidance or suggestions you may have.
Thank you for your work on this package.