Question: Error object and replacement value dimnames differ, when changing column names of assayData of ExpressionSet (although dimension is same)
0
gravatar for salamandra
29 days ago by
salamandra0
salamandra0 wrote:

I retrieved raw expression values from geo series supplementary files and did a normalization:

treatment1Samples <- c("GSM437316","GSM437286","GSM437305","GSM437297","GSM437277","GSM437282","GSM437269","GSM437302")
treatment2Samples <- c("GSM437147","GSM437138","GSM437311","GSM437218","GSM437114","GSM437234","GSM437165","GSM437299")
sampleNames <- c(treatment1Samples, treatment2Samples) 
rawData <- read.celfiles(pathSampleCELfiles)
normData <- rma(rawData)

As the resulting expressionset didn't have the complete sample metadata (phenoData) I retrieved it from GEO with:

seriesMatrixFilesList <- getGEO(GEOId, GSEMatrix=T) 

# following selects data of platform we want (there are two in this experiment):
for (i in 1:length(seriesMatrixFilesList)) { 
  if (as.character(unique(seriesMatrixFilesList[[i]]$platform_id)) == GPLId) {
    seriesMatrixFile <- seriesMatrixFilesList[[i]]
  }
}

pheno <- phenoData(seriesMatrixFile) # gets complete phenoData of the study
pheno <- pheno[sampleNames,] # pheno has metadata for all samples, so we select just the samples we want

and then added it to normalized expression values:

phenoData(normData) <- pheno

The rownames of phenodata has the samples names:

rownames(phenoData(normData))
[1] "GSM437316" "GSM437286" "GSM437305" "GSM437297" "GSM437277" "GSM437282" "GSM437269" "GSM437302" "GSM437147" "GSM437138" "GSM437311"
[12] "GSM437218" "GSM437114" "GSM437234" "GSM437165" "GSM437299"

But column names of assaydata are file names:

colnames(exprs(normData))
[1] "GSM437114.CEL.gz" "GSM437138.CEL.gz" "GSM437147.CEL.gz" "GSM437165.CEL.gz" "GSM437218.CEL.gz" "GSM437234.CEL.gz" "GSM437269.CEL.gz"
[8] "GSM437277.CEL.gz" "GSM437282.CEL.gz" "GSM437286.CEL.gz" "GSM437297.CEL.gz" "GSM437299.CEL.gz" "GSM437302.CEL.gz" "GSM437305.CEL.gz"
[15] "GSM437311.CEL.gz" "GSM437316.CEL.gz"

I tried to change assaydata column names to sample names:

newNames<- unlist(lapply(strsplit(colnames(exprs(normData)), '(\\.)|(_)'), function(x) x[1]))
newNames
[1] "GSM437114" "GSM437138" "GSM437147" "GSM437165" "GSM437218" "GSM437234" "GSM437269" "GSM437277" "GSM437282" "GSM437286" "GSM437297"
[12] "GSM437299" "GSM437302" "GSM437305" "GSM437311" "GSM437316"

colnames(exprs(normData)) <- newNames

And got this error:

Error in (function (od, vd)  : 
  object and replacement value dimnames differ

although number of element of newNames is same as of colnames(exprs(normData))

  1. How can I solve this?
  2. Also, does the order of column names in assay data has to be same as row names in phenoData?

Note: Above I din't do colnames(exprs(normData)) <- rownames(phenoData(normData)) cause wanted to keep assayData column names order, which was different from rownames order in phenoData

ADD COMMENTlink modified 29 days ago by James W. MacDonald51k • written 29 days ago by salamandra0
Answer: Error object and replacement value dimnames differ, when changing column names o
2
gravatar for James W. MacDonald
29 days ago by
United States
James W. MacDonald51k wrote:

You are recursing too deeply into the bowels of the ExpressionSet. Put a different way, in order for the colnames for the exprs slot of the ExpressionSet to be valid, they have to match up with the rownames of the phenoData slot. If you just change one, you get an error because they both have to be consistent. If you instead change the colnames of the ExpressionSet itself, the rownames of the phenoData slot get changed as well, and all is good.

> library(Biobase)
> data(sample.ExpressionSet)
> colnames(exprs(sample.ExpressionSet))
 [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" "K" "L" "M" "N" "O" "P" "Q" "R" "S"
[20] "T" "U" "V" "W" "X" "Y" "Z"
> colnames(sample.ExpressionSet)
 [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" "K" "L" "M" "N" "O" "P" "Q" "R" "S"
[20] "T" "U" "V" "W" "X" "Y" "Z"
> colnames(exprs(sample.ExpressionSet)) <- letters
Error in (function (od, vd)  : 
  object and replacement value dimnames differ
> colnames(sample.ExpressionSet) <- letters
> colnames(exprs(sample.ExpressionSet))
 [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "o" "p" "q" "r" "s"
[20] "t" "u" "v" "w" "x" "y" "z"
> colnames(sample.ExpressionSet)
 [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "o" "p" "q" "r" "s"
[20] "t" "u" "v" "w" "x" "y" "z"


ADD COMMENTlink written 29 days ago by James W. MacDonald51k

it was easier than I though. thank you!

ADD REPLYlink written 28 days ago by salamandra0
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 16.09
Traffic: 434 users visited in the last hour