Hello! I am doing a WGCNA and during the process of external trait matching I stumbled upon the problem, that in order to correlate the eigengenes I need numeric values. The problem is that I have categorical variables. Am I allowed to just convert them to numeric values, or does this mess up the statistic?

    traitData <- read.table("sampleinfo_comb.txt", header = T)

    Samples <- rownames(datExpr0)

    nGenes = ncol(datExpr0)
    nSamples = nrow(datExpr0)


    #In Order to correlate traits to the eigengenes, we need numeric values (Am I allowed to do it like this, or do I need another correlation method??)

    traitData$Treatment <- as.numeric(factor(traitData$Treatment, levels = c("no_challenge", "challenge")))
    traitData$Setting <- as.numeric(factor(traitData$Setting, levels = c("group", "isolation"))) #converts group=1 and isolation=2
    traitData$Caste <- as.numeric(factor(traitData$Caste, levels = c("worker", "queen"))) # converts worker=1 and queen=2

    #########

    module.trait.correlation = cor(mergedMEs, traitData, use = "p") #p for pearson correlation coefficient 
    module.trait.Pvalue = corPvalueStudent(module.trait.correlation, nSamples)

This code generated some pretty results for my data, I am just not sure whether it is the appropiate method for my categorical traits?

Thanks for any advice!