Question

remove X and Y chromosome genes in RNA-seq data using DESeq2 pipeline

0

Entering edit mode

anpham ▴ 60

@anpham-7402

Last seen 7.2 years ago

I'd like to remove the genes on the X and Y chromosomes from my human RNA-seq data before doing differential analysis using DESeq2. I've looked through the RNA-seq Workflow and DESeq2 manuals but didn't see this as an option. Any help in performing this step and still using the DESeq2 or RNA-seq Workflow pipeline would be much appreciated. Thanks.

deseq2 rnaseq • 5.6k views

ADD COMMENT • link updated 6.5 years ago by ta_awwad ▴ 10 • written 9.6 years ago by anpham ▴ 60

0

Entering edit mode

Thank you for your response. I followed the instruction and got an error message and not sure how to fix it. I used the UCSC hg19 to make the count matrix using GenomicAlignments. I copied the codes and the error message here. (By the way, I named my dds "dds1"). Thanks.

csvfile1 <- "table1.csv"
(sampleTable1 <- read.csv(csvfile1, row.names=1))

filenames1 <- file.path(paste0(sampleTable1$Run, ".bam"))

library("Rsamtools")
bamfiles1 <- BamFileList(filenames1, yieldSize=2000000)

biocLite("TxDb.Hsapiens.UCSC.hg19.knownGene")
source("http://bioconductor.org/biocLite.R")
library(TxDb.Hsapiens.UCSC.hg19.knownGene)

(genes <- transcriptsBy(TxDb.Hsapiens.UCSC.hg19.knownGene, by="gene"))

library("GenomicAlignments")
se1 <- summarizeOverlaps(features=genes, reads=bamfiles1, mode="Union", singleEnd=FALSE, ignore.strand=TRUE, fragments=TRUE)

library("DESeq2")
library("BiocParallel")
register(MulticoreParam(4))
dds1 <- DESeqDataSet(se, design = ~ gender + agecat3 + sfrace2 + RNAbatch + Site + PTSD_1mo_k)
dds1$PTSD_1mo_k <- relevel(dds1$PTSD_1mo_k, "control")

## subsetting dds1
seqnames(rowData(dds1))
dds1.sub <- dds1[ ! seqnames(rowData(dds1)) %in% c("chrX", "chrY"), ]

Error in x[i, , drop = FALSE] : invalid subscript type 'list'

ADD REPLY • link 9.6 years ago anpham ▴ 60

0

Entering edit mode

I forgot, I was thinking GRanges but you have a GRangesList.

[edit] see Martin's answer above.

ADD REPLY • link 9.6 years ago Michael Love 43k

0

Entering edit mode

Hi Mike,

I tried to do the same but it gave me this error message:

dds.sub <- dds[ ! seqnames(rowRanges(dds)) %in% c("X","Y"), ]
Error in (function (classes, fdef, mtable) :
unable to find an inherited method for function ‘NSBS’ for signature ‘"CompressedRleList"’

ADD REPLY • link 6.5 years ago ta_awwad ▴ 10

1

Entering edit mode

See Martins post below

ADD REPLY • link 6.5 years ago Michael Love 43k

0

Entering edit mode

thanks much mike, I tried it already but it doesn't seem working as the number of DE genes are still the same as if X and Y genes were not removed! .. I don't know what to do in order to make this correctly.

thanks

ADD REPLY • link 6.5 years ago ta_awwad ▴ 10

1

Entering edit mode

Are you using all()? I don’t see it in your code. Martin’s code shows how to properly subset if you have a GRangesList.

ADD REPLY • link 6.5 years ago Michael Love 43k

0

Entering edit mode

> seqnames(dds) %in% c("X", "Y")
RleList of length 52636

> !seqnames(dds) %in% c("X", "Y")
RleList of length 52636

ADD REPLY • link 6.5 years ago ta_awwad ▴ 10

score 0 · Answer 1 · 2015-05-01

0

Entering edit mode

Michael Love 43k

@mikelove

Last seen 1 day ago

United States

It's perfectly valid (statistically) to do this, but you might have to do some legwork. How did you make the count matrix? If you used summarizeOverlaps, the chromosomes are right at hand [edit: see Martin's comment for GRangesList]

seqnames(rowData(dds))

Or for the most recent release of Bioconductor (3.1):

seqnames(rowRanges(dds))

Then just subset the dds:

dds.sub <- dds[ ! seqnames(rowRanges(dds)) %in% c("chrX","chrY"), ]

ADD COMMENT • link 9.6 years ago • updated 5.7 years ago Michael Love 43k

2

Entering edit mode

SummarizedExperiment (and hence DESeqDataSet) 'knows' that it has ranges as rows, so seqnames(dds) is enough. For a GRangesList, I guess the assumption is that that all element ranges have the same chromosome, so dds[ all(!seqnames(dds) %in% c("chrX", "chrY")), ]. Unpacking a bit, seqnames() on a GRangesList returns an RleList of seqnames, one element of the list for each element of the GRangesList. %in% returns an RleList of logical values, again retaining the geometry. And all() is applied to each element of the list, returning a logical vector of the same length as the original GRangesList.

ADD REPLY • link 9.6 years ago Martin Morgan 25k