Feature Selection Bug?

Using a the feature selection criteria of 0.05 and log2Fc of 1, I get Total sig.genes of 1143 that are colored in orange. Yet, the sorted table --from lowest adjusted p value to the highest-- show only 1 gene that truly verifies those criteria.

I get the same output using networkanalyst3

Below is the R command history and I’ve attached a snapshot and the raw count data set.

PerformDataAnnot(NA, “mmu”, “count”, “symbol”, “mean”);
PlotDataBox(NA, “qc_boxplot_0_”, “72”, “png”);
PlotDataPCA(NA, “qc_pca_0_”,“72”, “png”, “NA”);
PlotLibSizeView(NA, “libsize_0_”,“72”,“png”,“NA”);
PlotDataDensity(NA, “qc_density_0_”,“72”, “png”, “NA”);
PerformExpressNormalization(NA, “logcount”, 15, 3, 5,“true”);
PlotDataBox(NA, “qc_norm_boxplot_0_”, “72”, “png”);
PlotDataPCA(NA, “qc_norm_pca_0_”,“72”, “png”, “NA”);
PlotDataDensity(NA, “qc_norm_density_0_”,“72”, “png”, “NA”);
PlotDataMeanStd(NA, “qc_norm_meanstd_0_”,“72”, “png”);
SetSelectedMetaInfo(NA, “CLASS”, “NA”, F)
SetupDesignMatrix(NA, “limma”);
PerformDEAnal(NA, “custom”, “TxO vs. CO”, “NA”, “intonly”);
SetupDesignMatrix(NA, “deseq2”);
PerformDEAnal(NA, “custom”, “TxO vs. CO”, “NA”, “intonly”);
All_CO_vs_All_TxO.txt (1.1 MB)

Hello Ahmed,

Thanks for letting me know, I will check it out.


Thank you so much for your help!

Hi Guangyan, are the network and express DEG reliable ? did you get a chance to examine the discrepancy i pointed out ? kindly advise. Many Thanks!

Hello Ahmed,

The issues is only related to that FeatureSelection page, I have fixed the bug but I can’t update the server just yet because I am in the middle of code rework. I hope to try update this week.


Thank you Guangyan,