Having problem on doing edgeR anlysis- cant create the DGE list through readbismark2DGE function.

Question

[Rabeya]

I am facing a problem on getting the DGElist from the function readbismark2dGE(). It didn't show any error. output of bismark2DGE only shows Hashing, counting. I am using google colab to get more system RAM. I am using R version R4.2.3 and BiocManager 3.16.

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("edgeR")

library(edgeR)
library(limma)

targets <- read.table("targets.txt", header=TRUE, stringsAsFactors=FALSE, fill=TRUE)

targets

s1 <- read.delim(file = "1760-read-1_bismark_bt2.bismark.cov.gz", header = FALSE, nrows = 6)
s1

    V1 V2        V3          V4 V5 V6
1 6 3121266 3121266 0.00 0 17
2 6 3121296 3121296 0.00 0 17
3 6 3179319 3179319 1.28 1 77
4 6 3180316 3180316 4.55 1 21
5 6 3182928 3182928 4.33 22 486
6 6 3182937 3182937 5.37 61 1074


library(readr)

BiocManager::install("methylKit", force = TRUE)

files <- targets$File

files
1760-read-1_bismark_bt2.bismark.cov.gz, 1760-read-1_bismark_bt2.bismark.cov.gz 
library(methylKit)

yall=readBismark2DGE(files, sample.names=targets$Sample)

The yall output only shows below output. I am using google colab to get more system RAM. I am using R version R4.2.3 and BiocManager 3.16.

Reading 1760-read-1_bismark_bt2.bismark.cov.gz 
Reading 1760-read-1_bismark_bt2.bismark.cov.gz 
Hashing ...
Collating counts ...

Answer 1

[Gordon Smyth]

The screen output you show is the correct output that the function is supposed to produce. There are no other messages sent to the screen. I am not clear what problem you are having. Are you saying that the function does not finish because of lack of memory? Otherwise, the function seems to have worked correctly. The DGEList is written to the yall object.

You code is installing and loading the methylKit package but none of the code requires methylKit.

[Rabeya]

Thank you very much for your valuable time. I am encountering difficulties in obtaining the DGEList and would greatly appreciate your assistance. It appears that the function fails to complete due to insufficient memory. Upon executing the relevant steps, an error message informs me that my session was terminated due to a shortage of RAM. Despite having upgraded to Google Colab Pro with a 25.5 GB RAM capacity, this issue persists. For your information, I am currently using R 4.2.3 and Bioconductor version 3.16.

I am referring to the edgeR manual to obtain differential methylation in my research work, with my two .cov files each sized at 197 Mb and 394 Mb, respectively. I would be most grateful if you could kindly advise me on how to resolve this memory issue. Thank you in advance for your kind attention to this matter.

I am following the below steps to get the DGEList.

library(edgeR)
library(limma)
library(readr)

targets <- read.table("targets.txt", header=TRUE, stringsAsFactors=FALSE, fill=TRUE)

s1 <- read.delim(file = "17-read_1_bismark_bt2.bismark.cov.gz", header = FALSE, nrows = 6)
files <- targets$File

yall <- readBismark2DGE(files, sample.names=NULL, readr=TRUE, verbose=TRUE)

yall <- readBismark2DGE(files, sample.names=targets$Sample)

[Gordon Smyth]

If you don't have enough memory to run the function, then there isn't anything can be done.

Have said that, I will note:

• 25GB RAM isn't a particularly large machine. In my Lab, even our laptops have more memory than that.
• Your cov files do seem pretty large.
• readBismark2DGE() is recommended for use with reduced representation bisulfite sequencing, not for analysis of whole genome data.

[Rabeya]

Thank you so much for your valuable time. I will increase the allocated memory in order to obtain the output of readBismark2DGE. My main goal is to obtain differential methylation, as I am currently in the initial phase of this process. Thanks again.