net$hub returning NULL

[Valentin-Bio-zz]

Hello, developer! thanks for this tool! I'm trying to make a camparison between networks and later establish differentially associated nodes using NetCoMi. THe point is when I run netAnalyze() over the MIcronet object with two networks I cannot plot the networks due to NULL hubs. I used SpiecEAsi adjacent networks for the analysis:

oral_NO_AR_net <- netConstruct(data = as.matrix(symBeta(getOptBeta(oral_AR_net)), mode = "maxabs"),
                               data2 = as.matrix(symBeta(getOptBeta(oral_NO_net)), mode = "maxabs"),
                               dataType = "condDependence", normMethod = "none", 
                               sparsMethod = "none", zeroMethod = "none", 
                               verbose = 3, seed = 1234)

then, when I run the netAnalyze() I got the following:

oral_NO_AR_comp <- netAnalyze(oral_NO_AR_net, centrLCC = FALSE, avDissIgnoreInf = TRUE,
                              clustMethod = "cluster_fast_greedy", 
                              hubPar = c("between","degree"), hubQuant = 0.95,
                              normDeg = TRUE, normBetw = TRUE)

oral_NO_AR_comp$hubs

**$hubs1 NULL

$hubs2 NULL**

[stefpeschel]

Hi Valentin, Could you please check whether your input has row and column names? E.g.: rownames(as.matrix(symBeta(getOptBeta(oral_AR_net)), mode = "maxabs")) colnames(as.matrix(symBeta(getOptBeta(oral_AR_net)), mode = "maxabs"))

They are important because they define the taxa, which are the network's nodes. Note that row and column names must be equal if association matrices are used as input.

I'll add a stopping criterion to catch the case that dim names are missing.

Best, Stefanie

[Valentin-Bio-zz]

Thanks so much for your help! it worked.

Best,

Valentín.