Question

minfi support of illumina Infinium methylationEPIC v2.0

0

Entering edit mode

luca.s ▴ 50

@lucas-24386

Last seen 4 months ago

Italy

Hi all. I have tried to process Infinium methylationEPIC v2.0 with minfi v1.44.0 but it did not work, apparently not recognizing the arrays: function "read.metharray.exp" results in "unknown" for the annotation. Is there a way to process these new arrays with minfi? Any suggestion/solution is greatly appreciated. Note that I am a newbe in using R. Thaks

methylationEPICv2.0 minfi • 8.3k views

ADD COMMENT • link updated 7 weeks ago by wl_xxz • 0 • written 20 months ago by luca.s ▴ 50

score 4 · Answer 1 · 2023-03-20

4

Entering edit mode

James W. MacDonald 67k

@james-w-macdonald-5106

Last seen 9 hours ago

United States

The annotation packages have not been generated by Bioconductor core, but a long time contributor has done so. You could install using BiocManager

library(BiocManager)
install("jokergoo/IlluminaHumanMethylationEPICv2manifest")
install("jokergoo/IlluminaHumanMethylationEPICv2anno.20a1.hg38")

Note that you may need to also install the remotes package for BiocManager to install from a GitHub repo.

ADD COMMENT • link 20 months ago James W. MacDonald 67k

0

Entering edit mode

Tank you James. All fine, but for preprocessIllumina, which apparently won't work

ADD REPLY • link 20 months ago luca.s ▴ 50

0

Entering edit mode

Thank you James for providing the annotation and manifest packages for methylationEPIC v2.0. I have installed and loaded them, also I have loaded minfi. I have a new batch of samples hybridized to the v2.0 array that I need to analyze. I have no problem to load the IDAT files and to create the object (which, by the way, I managed to create before installing v2.0 specific annotation and manifest), however it is still type "Unknown":

basedir <- "/1204_IDATS/"
targets <- read.metharray.sheet(basedir, pattern = "csv$", ignore.case = TRUE, recursive = FALSE, verbose = TRUE)
rgsetext <- read.metharray.exp(targets = targets, force = TRUE, extended = TRUE)
rgsetext

class: RGChannelSetExtended 
dim: 1105209 8 
metadata(0):
assays(5): Green Red GreenSD RedSD NBeads
rownames(1105209): 1600157 1600179 ... 99810982 99810990
rowData names(0):
colnames(8): 207059430150_R05C01 207059430150_R06C01 ... 207059430150_R03C01 207059430150_R04C01
colData names(45): sample_Name Sample_Name ... Basename filenames
Annotation
  array: Unknown
  annotation: Unknown

and therefore the functions getBeta(), preprocessRaw() and others like preprocessNoob() do not work:

betaval <- getBeta(rgsetext)
Error in .getManifestString(object@annotation) : 
  Cannot get Manifest object for an 'Unknown' array

msetext <- preprocessRaw(rgsetext)
    Error in .getManifestString(object@annotation) : 
      Cannot get Manifest object for an 'Unknown' array

msetext <- preprocessNoob(rgsetext)
    Error in .getManifestString(object@annotation) : 
      Cannot get Manifest object for an 'Unknown' array

Has anyone faced this issue? thanks a lot in advance

ADD REPLY • link 20 months ago mariagallardo ▴ 10

1

Entering edit mode

setting the right array type and annotation will do the work:

rgsetext@annotation <- c(array = "IlluminaHumanMethylationEPICv2", annotation = "20a1.hg38")
rgsetext

class: RGChannelSetExtended 
dim: 1105209 8 
metadata(0):
assays(5): Green Red GreenSD RedSD NBeads
rownames(1105209): 1600157 1600179 ... 99810982 99810990
rowData names(0):
colnames(8): 207059430150_R05C01 207059430150_R06C01 ... 207059430150_R03C01 207059430150_R04C01
colData names(45): sample_Name Sample_Name ... Basename filenames
Annotation
  array: IlluminaHumanMethylationEPICv2
  annotation: 20a1.hg38

ADD REPLY • link 20 months ago mariagallardo ▴ 10

1

Entering edit mode

For me, your solution but using c(array = "IlluminaHumanMethylationEPICv2", annotation = "ilm10b2.hg19") worked fine. Also see https://github.com/hansenlab/minfi/issues/97

ADD REPLY • link 13 months ago jay ▴ 10

0

Entering edit mode

Hello James! Is there an annotation package for the latest EPICv2 A2 sub-version? If not, I am happy to make one by modifying the existing package. p.s.: never mind! i see A2 is only a minor change of adding 2 columns of infrmation from A1

ADD REPLY • link 7 weeks ago wl_xxz • 0