search

Oshlack / missMethyl

Bioconductor package for analysis of methylation data from Illumina's Infinium HumanMethylation arrays.
10 stars 4 forks source link

A question about `gometh` #1

Closed QianhuiWan closed 5 years ago

QianhuiWan commented 6 years ago

Hi Oshlack, I updated missMethyl package last week, and when I rerun my analysis I found my results fromgometh and gsameth all changed. Do you know why this happen? Do I need to change back to the old version? Thank you so much. My codes are as follows:

c2_reactome<-read.gmt("/home/qianhui/DNAme/data/GSEA_MsigDB/c2.cp.reactome.v6.2.entrez.gmt") 

c2_reactomeList <- split(c2_reactome$gene, c2_reactome$ont) 

gsaMeth_DMRprobesReactome <- gsameth(sig.cpg = ProbesInDMRs.names, all.cpg = allProbesInDMRs.names, collection = c2_reactomeList, array.type = "EPIC", prior.prob = TRUE, ann=ann850k)

gsaMeth_DMRprobesReactome[gsaMeth_DMRprobesReactome[,4]<0.05,] 

topGSA(gsaMeth_DMRprobesReactome, number = 10)
JovMaksimovic commented 6 years ago

Hi, I have updated missMethyl to now use the newest Illumina annotation (IlluminaHumanMethylationEPICanno.ilm10b4.hg19) by default. You can either revert to using the Bioconductor release version of missMethyl, which still uses the older annotation, or you can pass the older annotation to the gometh function directly:

# load older version of the Illumina annotation 
library(IlluminaHumanMethylation450kanno.ilmn12.hg19)
# get older version of the annotation
ann <- getAnnotation(IlluminaHumanMethylation450kanno.ilmn12.hg19)
# pass old annotation to gometh using the "anno" parameter.
gst <- gometh(sig.cpg = sigcpgs, all.cpg = allcpgs, collection = "GO", prior.prob = TRUE, anno = ann)
QianhuiWan commented 6 years ago

Thank you so much!