DESeqDataSetFromHTSeqCount function taking long time and utilizing more RAM

Question

[Ranan Jyoti Sarma]

I am using DESeq2 for differential gene expression. I have done alignment using STAR aligner and GRCh38 reference genome. I have generated count files using HTSeq. I have 3 controls and 4 cases and the size of each count file varies from 8 GB to 11 GB which makes to total 73 GB of data (7 Count files).

ddsHTSeq <- DESeqDataSetFromHTSeqCount(sampleTable = sampleTable, directory = directory, design = ~ condition)

directory is where I kept the count files.

This command is taking more than 12 hours and consumed memory of 46 GB. I have a maximum memory limit of 64 GB in my System.

Could anyone help how to solve this problem?

[GenoMax]

size of each count file varies from 8 GB to 11 GB

That is not likely to be correct. I assume these are original BAM alignment files. Even with 10 samples a count matrix file for featureCounts for mouse samples is < 30 MB.

[ATpoint]

Something is not right here. Please share all relevant command lines. Is this normal RNA-seq?

[Ranan Jyoti Sarma]

Yes. This is a normal RNA-seq.

directory="~/Desktop/deseq2/"
    setwd(directory)
    sampleNames<- c("AN_d83", "AN_d85","AN_d86", "Tum_t83","Tum_t84","Tum_t85","Tum_t86")
    sampleCondition<- c("Adjacent Normal","Adjacent Normal","Adjacent Normal","Tumor","Tumor","Tumor","Tumor")
    sampleTable <- data.frame(sampleName = sampleNames, fileName = sampleFiles, condition = sampleCondition)
    treatments<- c("Adjacent Normal","Tumor")
    library("DESeq2")
    ddsHTSeq <- DESeqDataSetFromHTSeqCount(sampleTable = sampleTable, directory = directory, design = ~ condition)

[ATpoint]

As genomax says these files are way to big for count matrices. The loading process should take a couple of seconds actually and should be possible on any standard laptop. Check that file paths are correct and really direct to the count matrices, not to the bam files.

[Ranan Jyoti Sarma]

Thanks, Problem is solved.

[ATpoint]

Please add the solution to the answer field if others encounter the same problem.