Problem in analysis of Ballgown output by R (why q value are same in result ? )

Question

[amitunited0532]

Hello,

I have performed an RNAseq experiment, for which I used HISAT2 for the alingment, Stringtie for the assembly and the R package Ballgown for the Differential Expression (DE) analysis. After generation of differential expression table, I found q value is mostly same for all genes which is given below.

Command which I run:

results_genes = stattest(bacteria_filter, feature = "gene", covariate = "sample", getFC = TRUE, meas = "FPKM") 
results_genes = merge(results_genes,bacteria_gene_names,by.x=c("id"),by.y=c("gene_id"))

differential_genes = subset(results_genes, results_genes$pval<0.05)
write.csv(differential_genes, "differential_genes.csv", row.names = FALSE)

Output:

qval

0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586 0.999988586

Thank you

[ATpoint]

Please use the code rather than the blockquote buttom to indicate code :)

[russhh]

you probably don't have any significant hits even at non-FDR-adjusted levels. FDR-adjustment does lead to chunking quite often, though it's rarely a concern: if you need to rank by reproducibility use the untransformed levels, if you need to test at a certain threshold, use the FDR-adjusted levels