dmrseq fails while processing chromosomes/contigs
jckhearn opened this issue · 2 comments
Hi @kdkorthauer,
I am running dmrseq on WGBS of a sparsely methylated genome and it is failing with the following error:
Error in
[[<-.data.frame(
tmp, i, value = c(1L, 1L, 1L, 1L, 1L, 1L, : replacement has 480 rows, data has 240 Calls: dmrseq ... model.matrix -> model.matrix.default -> [[<- -> [[<-.data.frame
The exact contig at which failure occurs varies according to the parameters applied to dmrseq
but also crops up. I have filtered out contigs less than 1000bp.
The experiment is a complex design with 48 replicates total and fails when testing 'Food' against 'Age' and 'Strain' as adjustCovariates
:
testFoodCovariate<-"Food"
matchStrainAgeCovariate<-c("Strain","Age")
Food <- dmrseq(bs= bismarkBSseq.filtered, cutoff = 0.05, testCovariate=testFoodCovariate,adjustCovariate=matchStrainAgeCovariate,maxPerms = 10)
Food | Age | Strain |
---|---|---|
H | O | C32 |
H | O | C32 |
H | O | C32 |
H | O | C32 |
H | O | C32 |
H | O | C32 |
L | O | C32 |
L | O | C32 |
L | O | C32 |
L | O | C32 |
L | O | C32 |
L | O | C32 |
H | Y | C32 |
H | Y | C32 |
H | Y | C32 |
H | Y | C32 |
H | Y | C32 |
H | Y | C32 |
L | Y | C32 |
L | Y | C32 |
L | Y | C32 |
L | Y | C32 |
L | Y | C32 |
L | Y | C32 |
H | O | KA53 |
H | O | KA53 |
H | O | KA53 |
H | O | KA53 |
H | O | KA53 |
H | O | KA53 |
L | O | KA53 |
L | O | KA53 |
L | O | KA53 |
L | O | KA53 |
L | O | KA53 |
L | O | KA53 |
H | Y | KA53 |
H | Y | KA53 |
H | Y | KA53 |
H | Y | KA53 |
H | Y | KA53 |
H | Y | KA53 |
L | Y | KA53 |
L | Y | KA53 |
L | Y | KA53 |
L | Y | KA53 |
L | Y | KA53 |
L | Y | KA53 |
Bioconductor packages are up to date. Why do you think this error is occurring?
Thanks for any help!
Hi @jckhearn,
Would you mind sending me a small subset of your bsseq object (as an rds file) that throws the error so that I can see what's going on? Just a single chromosome or contig is sufficient, as long as it throws the error.
You can send to keegan@jimmy.harvard.edu if you'd prefer.
Thank you!
Best,
Keegan
Will do over the next couple of days!
Thanks,
Jack