DESeq error of my data?

Question

[PK]

I am doing RNASeq analysis my pipe line is bowtie>tophat>htseq>DESeq. I have two conditions and two replicates of each condition. R version 3.3.2

library(DESeq2)

countsTable <- read.delim("crassa_merged_A_G.txt")

countsTable <- countsTable[,-1]

conditions <- factor(c("arabinose","arabinose","glucose","glucose"))

library(DESeq)

countDataSet <- newCountDataSet(countsTable, conditions)

countDataSet <- estimateSizeFactors(countDataSet)

sizeFactors(countDataSet)

head(counts(countDataSet))

head(counts(countDataSet,normalized=TRUE))

countDataSet <-estimateDispersions(countDataSet)

fpkm(countDataSet, robust = TRUE)

plotDispEsts(countDataSet)

dev.off()

plotDispEsts <- function(countDataSet){ plot(rowMeans( counts( countDataSet, normalized=TRUE ) ), fitInfo(countDataSet) $perGeneDispEsts, pch = '.', log="xy", ylab ="dispersion", xlab ="mean of normalized counts") xg = 10^seq( -.5, 5, length.out=300 ) lines( xg, fitInfo(countDataSet)$dispFun( xg ), col="red" )}

Error: unexpected symbol in "plotDispEsts <- function(countDataSet){plot(rowMeans(counts(countDataSet,normalized=TRUE) ),fitInfo(countDataSet)$perGeneDispEsts, pch='.',log="xy", ylab="dispersion",xlab="mean of normalized "

I am getting an error like that. please tell me

[Carlo Yague]

Why are you trying to rewrite the plotDistEsts function ? That's weird.

Anyway, you forgot the newline characters (or alternatively the semicolons). This should work :

plotDispEsts2 <- function(countDataSet){
    plot(rowMeans( counts( countDataSet, normalized=TRUE ) ), fitInfo(countDataSet) $perGeneDispEsts, pch = '.', log="xy", ylab ="dispersion", xlab ="mean of normalized counts")
    xg = 10^seq( -.5, 5, length.out=300 )
    lines( xg, fitInfo(countDataSet)$dispFun( xg ), col="red" )
}

[WouterDeCoster]

Why are you trying to rewrite the plotDistEsts function

I doubt whether that's what he really wants to do.

[PK]

Thanks for your reply and valuable suggestion. It works now.

[Matteo Schiavinato]

Not a solution, but a suggestion: since your pipeline includes htseq as a counting algorithm, why don't you simplify your pipeline using the built-in function for htseq in DESeq2?

DESeqDataSetFromHTSeqCount(sampleTable, directory = ".", design, ignoreRank = FALSE, ...)

This will create a seqExpressionSet already perfectly set for the DESeq() function.

Make sure to post the complete error message.

I associate myself in this request.

[WouterDeCoster]

Please use code formatting using the 101010 button. Select your code block, then click 101010.

Make sure to post the complete error message.