minfi differential methylation analysis errors
1
0
Entering edit mode
@5c7c40e5
Last seen 14 months ago
United States

Hi everyone,

I'm still pretty much an R novice and I'm trying to run an analysis on some EPIC array data I've been given. I am trying to analyse this to understand whether there is differential methylation between 3 different conditions. I have been following the minfi tutorial (https://www.bioconductor.org/help/course-materials/2015/BioC2015/methylation450k.html#quality-control) and trying to adapt the script to make it work with my data. In the tutorial that I'm using, they're analysing how methylation is altered by sex or age, and I need to look at how methylation is altered by insulin resistance. However, I'm getting an error and I'm not sure how to fix it. I think that it's probably related to the sample sheet, or that I'm calling the wrong column. Below is a screenshot of the sample sheet:

enter image description here

I am running the package minfi (version 1.46.0) in RStudio. The R version I have is 4.3.1 and RStudio is 2022.07.1, build 554.

#Load packages
library(minfi)
library(sva)

#Set working directory
setwd("/Users/matthewsinton/Desktop/Manchester Postdoc/EPIC Array Data/IDAT")

#Set location for IDAT files
dataDirectory <- "/Users/matthewsinton/Desktop/Manchester Postdoc/EPIC Array Data/IDAT"
# list the files
list.files(dataDirectory, recursive = TRUE)

#Set location for sample sheet
targets <- read.metharray.sheet(dataDirectory, pattern="SampleSheet.csv")
targets

# read in the raw data from the IDAT files; warnings can be ignored.
RGSet <- read.metharray.exp(targets=targets)

# Get an overview of the data
phenoData <- pData(RGSet)
phenoData[,1:6]
manifest <- getManifest(RGSet)
manifest
head(getProbeInfo(manifest))

#MethylSet and RatioSet. A MethylSet objects contains only the methylated and unmethylated signals
MSet <- preprocessRaw(RGSet) 
MSet
head(getMeth(MSet)[,1:3])
head(getUnmeth(MSet)[,1:3])
RSet <- ratioConvert(MSet, what = "both", keepCN = TRUE)
RSet
beta <- getBeta(RSet)

#Genomic Ratio Set. The function mapToGenome applied to a RatioSet object will add genomic coordinates to each probe together with some additional annotation information. 
GRset <- mapToGenome(RSet)
GRset
beta <- getBeta(GRset)
M <- getM(GRset)
CN <- getCN(GRset)
sampleNames <- sampleNames(GRset)
probeNames <- featureNames(GRset)
pheno <- pData(GRset)
#To return the probe locations as a GenomicRanges objects, one can use the accessor granges:
gr <- granges(GRset)
head(gr, n= 3)
annotation <- getAnnotation(GRset)
names(annotation)

#Run QC on the data
head(getMeth(MSet)[,1:3])
head(getUnmeth(MSet)[,1:3])
qc <- getQC(MSet)
head(qc)
plotQC(qc)
densityPlot(MSet, sampGroups = phenoData$Sample_Group)
densityBeanPlot(MSet, sampGroups = phenoData$Sample_Group)
controlStripPlot(RGSet, controls="BISULFITE CONVERSION II")
qcReport(RGSet, pdf= "/Users/matthewsinton/Desktop/Manchester Postdoc/EPIC Array Data/qcReport.pdf")

#Preprocessing and normalisation
MSet.illumina <- preprocessIllumina(RGSet, bg.correct = TRUE,
                                    normalize = "controls")
MSet.swan <- preprocessSWAN(RGSet)
GRset.quantile <- preprocessQuantile(RGSet, fixOutliers = TRUE,
                                     removeBadSamples = TRUE, badSampleCutoff = 10.5,
                                     quantileNormalize = TRUE, stratified = TRUE, 
                                     mergeManifest = FALSE, sex = NULL)
MSet.noob <- preprocessNoob(RGSet)
GRset.funnorm <- preprocessFunnorm(RGSet)
snps <- getSnpInfo(GRset)
head(snps,10)
GRset <- addSnpInfo(GRset)
GRset <- dropLociWithSnps(GRset, snps=c("SBE","CpG"), maf=0)

#Identifying DMPs and DMRs
beta <- getBeta(GRset.funnorm)
ID  <- pData(GRset.funnorm)$ID
dmp <- dmpFinder(beta, pheno = ID  , type = "continuous")
head(dmp)

The error I get is:

Error in smooth.spline(lambda, vec.p0, w = ncs.weights, df = 3) : 
  missing or infinite values in inputs are not allowed

Any help solving this issue would be super helpful. As I say, I'm still a novice with R, so very happy to provide any info I've missed or to clarify anything.

minfi • 1.1k views
ADD COMMENT
0
Entering edit mode
@james-w-macdonald-5106
Last seen 2 days ago
United States

You shouldn't use conventional linear regression with beta values (which are bounded on [0,1] and definitely do not meet the assumptions for normal-based linear regression). Instead you should use the M-values. So something like

dmp <- dmpFinder(getM(GRset.funnorm), pheno = ID, type = "continuous")
ADD COMMENT
0
Entering edit mode

Hi, thanks so much for helping here. Unfortunately this also didn't work, and I got the following error when I ran that bit of script:

Error in smooth.spline(lambda, vec.p0, w = ncs.weights, df = 3) : 
  missing or infinite values in inputs are not allowed
In addition: Warning message:
Partial NA coefficients for 9 probe(s)

Do you have any ideas why that might be?

Thanks!

ADD REPLY

Login before adding your answer.

Traffic: 511 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6