Search
Question: minfi MethylationEPIC errors
1
gravatar for igor
20 months ago by
igor20
United States
igor20 wrote:

I am using the minfi devel version 1.17.6 because it has been updated to support the new MethylationEPIC. I downloaded IlluminaHumanMethylationEPICmanifest_0.2.0.tar.gz and IlluminaHumanMethylationEPICanno.ilmn10b.hg19_0.2.0.tar.gz
from https://bitbucket.org/hansenlab/illumina_epic. I am trying to process MethylationEPIC arrays. The read.metharray.exp() function works. However, many downstream functions like qcReport()densityPlot(), and getBeta() give me an error:

Error in packageVersion("IlluminaHumanMethylation450kmanifest") : 
  package ‘IlluminaHumanMethylation450kmanifest’ not found

Is that a problem that it's asking for the wrong manifest? Should I just install and use that library?

 

0
gravatar for Kasper Daniel Hansen
20 months ago by
United States
Kasper Daniel Hansen6.3k wrote:
It sounds like a bug, but I cannot reproduce this. Could you give a full transcript of what you're doing and could you show the output of annotation(RGSET) dim(RGSET) where RGSET is the output of read.metharray.exp(). Best, Kasper On Thu, Mar 17, 2016 at 11:14 PM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Question: > minfi MethylationEPIC errors <https: support.bioconductor.org="" p="" 79795=""/>: > > I am using the minfi devel version 1.17.6 because it has been updated to > support the new MethylationEPIC. I downloaded > IlluminaHumanMethylationEPICmanifest_0.2.0.tar.gz and > IlluminaHumanMethylationEPICanno.ilmn10b.hg19_0.2.0.tar.gz > from https://bitbucket.org/hansenlab/illumina_epic. I am trying to > process MethylationEPIC arrays. The read.metharray.exp() function works. > However, many downstream functions like qcReport(), densityPlot(), and > getBeta() give me an error: > > Error in packageVersion("IlluminaHumanMethylation450kmanifest") : > package ‘IlluminaHumanMethylation450kmanifest’ not found > > Is that a problem that it's asking for the wrong manifest? Should I just > install and use that library? > > > > ------------------------------ > > Post tags: minfi, methylation, IlluminaHumanMethylationEPICmanifest, > IlluminaHumanMethylationEPICanno.ilmn10b.hg19 > > You may reply via email or visit minfi MethylationEPIC errors >
ADD COMMENTlink written 20 months ago by Kasper Daniel Hansen6.3k

This is my code:

library(minfi)
library(IlluminaHumanMethylationEPICmanifest)
library(IlluminaHumanMethylationEPICanno.ilmn10b.hg19)

idatPath = "./idats"

targets = read.csv("samples.csv", strip.white=T, stringsAsFactors=F)

rawSet = read.metharray.exp(base=idatPath, targets=targets, verbose=T)

qcReport(rawSet, sampNames=pData(rawSet)$RD, sampGroups=pData(rawSet)$Condition, pdf="qcReport.pdf")

betaRaw = getBeta(rawSet)

normSet = preprocessNoob(rawSet)

And this is the info you asked for (which I think looks fine):

> annotation(rawSet)
                         array                     annotation 
"IlluminaHumanMethylationEPIC"                 "ilmn10b.hg19" 
> dim(rawSet)
Features  Samples 
 1052641       16 

Let me know if you need anything else.

 

ADD REPLYlink written 20 months ago by igor20

Quick update. I just installed "IlluminaHumanMethylation450kmanifest" library and now everything seems to be working. Even though there are no errors, I am still concerned that minfi is using the wrong manifest somehow.

ADD REPLYlink written 20 months ago by igor20
Ok, I found the bug and this should not affect anything with the results. In the preprocessRaw function, I record the version of the manifest file used to produce the object. Here, the right manifest file gets used for the computation, but when I write the version string I have hard coded packageVersion("IlluminaHumanMethylation450kmanifest"). This call fails when the package is not installed. But it only affects a cosmetic thing. You can see it (now that you can run preprocessRaw) in the output, I get something like MethylSet (storageMode: lockedEnvironment) assayData: 866836 features, 3 samples element names: Meth, Unmeth An object of class 'AnnotatedDataFrame' sampleNames: 200144450018_R04C01 200144450019_R07C01 200144450021_R05C01 varLabels: Sample_Name Sample_Well ... filenames (9 total) varMetadata: labelDescription Annotation array: IlluminaHumanMethylationEPIC annotation: ilmn10b.hg19 Preprocessing Method: Raw (no normalization or bg correction) minfi version: 1.17.6 Manifest version: 0.4.0 Look at how the manifest version is 0.4.0 and not 0.2.0. Thanks for the bug report, I will fix this soon, and I will look for other cases of hard coding. Btw., I see Illumina has released a new annotation file for the EPIC array. I'll update the packages soon. Best, Kasper On Thu, Mar 17, 2016 at 11:48 PM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Comment: minfi > MethylationEPIC errors <https: support.bioconductor.org="" p="" 79795="" #79798="">: > > Quick update. I just installed "IlluminaHumanMethylation450kmanifest" > library and now everything seems to be working. Even though there are no > errors, I am still concerned that minfi is using the wrong manifest somehow. > > ------------------------------ > > Post tags: minfi, methylation, IlluminaHumanMethylationEPICmanifest, > IlluminaHumanMethylationEPICanno.ilmn10b.hg19 > > You may reply via email or visit > C: minfi MethylationEPIC errors >
ADD REPLYlink written 20 months ago by Kasper Daniel Hansen6.3k

Thanks for clarifying. I think I am okay then. One minor issue I have is my MethylSet is missing preprocessing info (but I saw that with the stable minfi and 450K arrays as well, so that may be due to how preprocessNoob() works):

> normSet = preprocessNoob(rawSet)
[preprocessNoob] Using sample number 5 as reference level...
> class(normSet)
[1] "MethylSet"
attr(,"package")
[1] "minfi"
> normSet
MethylSet (storageMode: lockedEnvironment)
assayData: 866836 features, 16 samples 
  element names: Meth, Unmeth 
An object of class 'AnnotatedDataFrame'
  sampleNames: 200498360077_R01C01 200498360077_R02C01 ... 200498360111_R08C01 (16 total)
  varLabels: Basename RD ... filenames (7 total)
  varMetadata: labelDescription
Annotation
  array: IlluminaHumanMethylationEPIC
  annotation: ilmn10b.hg19
Preprocessing
  Method: NA
  minfi version: NA
  Manifest version: NA

 

ADD REPLYlink written 20 months ago by igor20
yeah, that should be fixed as well. I think it is a good idea to record the preprocessing method inside a [Genomic][Methyl|Ratio]Set, but I never got the implementation to be right. That doesn't excuse missing this, of course. On Fri, Mar 18, 2016 at 12:10 AM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Comment: minfi > MethylationEPIC errors <https: support.bioconductor.org="" p="" 79795="" #79800="">: > > Thanks for clarifying. I think I am okay then. One minor issue I have is > my MethylSet is missing preprocessing info (but I saw that with the stable > minfi and 450K arrays as well, so that may be due to how preprocessNoob() > works): > > > normSet = preprocessNoob(rawSet) > [preprocessNoob] Using sample number 5 as reference level... > > class(normSet) > [1] "MethylSet" > attr(,"package") > [1] "minfi" > > normSet > MethylSet (storageMode: lockedEnvironment) > assayData: 866836 features, 16 samples > element names: Meth, Unmeth > An object of class 'AnnotatedDataFrame' > sampleNames: 200498360077_R01C01 200498360077_R02C01 ... 200498360111_R08C01 (16 total) > varLabels: Basename RD ... filenames (7 total) > varMetadata: labelDescription > Annotation > array: IlluminaHumanMethylationEPIC > annotation: ilmn10b.hg19 > Preprocessing > Method: NA > minfi version: NA > Manifest version: NA > > > > ------------------------------ > > Post tags: minfi, methylation, IlluminaHumanMethylationEPICmanifest, > IlluminaHumanMethylationEPICanno.ilmn10b.hg19 > > You may reply via email or visit > C: minfi MethylationEPIC errors >
ADD REPLYlink written 20 months ago by Kasper Daniel Hansen6.3k

Thanks again. And in case you wondering, if I use a different preprocess function, that info shows up:

> q = preprocessQuantile(rawSet)
[preprocessQuantile] Mapping to genome.
[preprocessQuantile] Fixing outliers.
[preprocessQuantile] Quantile normalizing.
> q
class: GenomicRatioSet 
dim: 866836 16 
metadata(0):
assays(2): M CN
rownames(866836): cg14817997 cg26928153 ... cg07587934 cg16855331
rowData names(0):
colnames(16): 200498360077_R01C01 200498360077_R02C01 ... 200498360111_R07C01 200498360111_R08C01
colData names(8): Basename RD ... filenames predictedSex
Annotation
  array: IlluminaHumanMethylationEPIC
  annotation: ilmn10b.hg19
Preprocessing
  Method: Raw (no normalization or bg correction)
  minfi version: 1.17.6
  Manifest version: 0.4.0

Will the updates be posted to the Bioconductor devel version soon? I have no idea how that whole process works.

ADD REPLYlink written 20 months ago by igor20
When I fix it in Bioconductor svn it will usually take 24h-36h to show up on the build servers (meaning you can get it using biocLite()), provided no mistake is made (by mistake, here I am referring to me making a fix which results in the package not passing its own checks, which should not happen in a perfect world, but does happen occasionally). Kasper On Fri, Mar 18, 2016 at 12:24 AM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Comment: minfi > MethylationEPIC errors <https: support.bioconductor.org="" p="" 79795="" #79802="">: > > Thanks again. And in case you wondering, if I use a different preprocess > function, that info shows up: > > > q = preprocessQuantile(rawSet) > [preprocessQuantile] Mapping to genome. > [preprocessQuantile] Fixing outliers. > [preprocessQuantile] Quantile normalizing. > > q > class: GenomicRatioSet > dim: 866836 16 > metadata(0): > assays(2): M CN > rownames(866836): cg14817997 cg26928153 ... cg07587934 cg16855331 > rowData names(0): > colnames(16): 200498360077_R01C01 200498360077_R02C01 ... 200498360111_R07C01 200498360111_R08C01 > colData names(8): Basename RD ... filenames predictedSex > Annotation > array: IlluminaHumanMethylationEPIC > annotation: ilmn10b.hg19 > Preprocessing > Method: Raw (no normalization or bg correction) > minfi version: 1.17.6 > Manifest version: 0.4.0 > > Will the updates be posted to the Bioconductor devel version soon? I have > no idea how that whole process works. > > ------------------------------ > > Post tags: minfi, methylation, IlluminaHumanMethylationEPICmanifest, > IlluminaHumanMethylationEPICanno.ilmn10b.hg19 > > You may reply via email or visit > C: minfi MethylationEPIC errors >
ADD REPLYlink written 20 months ago by Kasper Daniel Hansen6.3k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 115 users visited in the last hour