WGCNA soft-threshold power selection
2
0
Entering edit mode
3.2 years ago
mahnazkiani ▴ 50

Hello,

I have a question about the results of WGCNA for selecting soft-thresholding power, all power numbers are below the line in the scale independence graph, and I don't know which power I need to select for next step.I couldn't insert the picture to be more clear, but code that I used was as follows:

# choose soft-thresholding power
powers = c(c(1:10), seq(from = 12, to=20, by=2))
sft = pickSoftThreshold(datExpr, powerVector = powers, verbose = 5)
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")
abline(h=0.9,col="red")

plot(sft$fitIndices[,1], sft$fitIndices[,5],
xlab="Soft Threshold (power)",ylab="Mean Connectivity", type="n",
main = paste("Mean connectivity"))
text(sft$fitIndices[,1], sft$fitIndices[,5], labels=powers, cex=cex1,col="red")

for (i in 1:dim(datExpr)[2]) {
datExpr[,i] <- c(as.double(datExpr[[i]]))
}

RNA-Seq • 5.0k views
0
Entering edit mode
0
Entering edit mode

Thank you for sharing the link

0
Entering edit mode

Why are you using -sign(sft$fitIndices[,3])*sft$fitIndices[,2], here?

0
Entering edit mode
3.2 years ago

0
Entering edit mode

0
Entering edit mode

Choose 6, for now. I am also going to hypothesise that your dataset is rather 'flat'. If you plot a histogram (hist()), how does it appear? Which type of input data re you using? - please elaborate on the processing steps.

0
Entering edit mode

I am using log2(x+1) normalized total counts RNa-seq data.

0
Entering edit mode

Hi Kevin,

Although it's late, thank you for your feedback. You mentioned that the dataset is possibly flatted? did you say due to the negative signed R2? please kindly tell me how to solve this issue?

0
Entering edit mode

This is covered in the FAQ of the WCGNA webpage https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/faq.html See #6: "I can't get a good scale-free topology index no matter how high I set the soft-thresholding power."

It kind of depends on your data and how many samples you have