WGCNA: very low genes assigned to the grey Module

Hello Im trying to see the function of unknown ncRNA during process by rna seq of 14 samples at different time points as following:-

dds <- DESeqDataSetFromMatrix(countData = data, colData = sample.info, design = ~time) keep <- rowSums(counts(dds)) >= 10

dds <- dds[keep,]

dds <- estimateSizeFactors(dds)

dds_lrt <- DESeq(dds, test="LRT", reduced= ~ 1) resultsNames(dds_lrt) datExpr0 <- assay(dds_lrt) gsg <- goodSamplesGenes(datExpr0, verbose = 5) gsg$allOK TRUE

vsd = getVarianceStabilizedData(dds_lrt)

vsd2 <- t(vsd)

datExpr<- vsd2

powers = c(c(1:20), seq(from = 12, to=20, by=2))

sft = pickSoftThreshold(datExpr, powerVector = powers, verbose = 5, dataIsExpr = T)

sizeGrWindow(9, 5) par(mfrow = c(1,2)); cex1 = 0.9; plot(sft$fitIndices[,1], -sign(sft$fitIndices[,3])sft$fitIndices[,2], xlab="Soft Threshold (power)",ylab="Scale Free Topology Model Fit,signed R^2",type="n", main = paste("Scale independence")); text(sft$fitIndices[,1], -sign(sft$fitIndices[,3])sft$fitIndices[,2], labels=powers,cex=cex1,col="red"); softPower = 7

adjacency = adjacency(datExpr, power = softPower);

TOM = TOMsimilarity(adjacency);

dissTOM = 1-TOM

collectGarbage()

as.dist(TOM)

geneTree = hclust(as.dist(dissTOM), method = "average");

minModuleSize = 30;

dynamicMods = cutreeDynamic(dendro = geneTree, distM = dissTOM, deepSplit = 2, pamRespectsDendro = FALSE, minClusterSize = minModuleSize);

table(dynamicMods)

dynamicColors = labels2colors(dynamicMods)

table(dynamicColors)

dynamicColors

black          blue         brown          cyan     darkgreen 
      380           927           714           132            83 
 darkgrey    darkorange       darkred darkturquoise         green 
       71            69            86            71           527 
 greenyellow          grey        grey60     lightcyan    lightgreen 
      156             1           122           128           105 
 lightyellow       magenta  midnightblue        orange paleturquoise 
       92           226           131            70            52 
     pink        purple           red     royalblue   saddlebrown 
      229           197           520            88            65 
   salmon       skyblue     steelblue           tan     turquoise 
      150            67            53           152           931 
   violet         white        yellow 
       48            67           554

all the genes showed to be in clusteres and only 2 genes are in grey module

while looking into the expression pattern of the genes inside each module they dont seems to share similar expression pattern what is the mistake I have done

rna-seq sequencing ncrna wgcna

• 2,772 views • link

updated 5.5 years ago by Ruben •

• written 5.5 years ago by Shero •

And what's the question exactly?

• 0 views • link

written 5.5 years ago by WouterDeCoster

all the genes showed to be in clusteres and only 2 genes are in grey module

while looking into the expression pattern of the genes inside each module they don't seems to share similar expression pattern what is the mistake I have done

• 0 views • link

written 5.5 years ago by Shero •

1 answer

Without doing all the work ourselves it would be very difficult to answer this question. Obviously, you can play around with the minModulesize parameter, the soft-power you use or the way you cut the dynamic modules. Any of these factors will influence the final results. I for one would flip the question, why are the genes in the grey module not assigned to another label? Hence, perhaps cut your modules differently so the grey module disappears.

One thing I can point as well: WGCNA should be run over data that is unbiased by your experimental design. Of course DESeq2 and vst do take the experimental design in consideration, but the WGCNA authors recommend it themselves on their website. However, if genes don't follow a similar pattern it does not mean that they didn't have similar vst expression values. Simply put, expression patterns informed by experimental design are not a good metric to look at the validity of clustering via WGCNA. The clustering tree would be a good place to start and see if grey's genes couldn't be assigned elsewhere.

• 0 views • link

written 5.5 years ago by Ruben •