hello....

i am trying ti find out differential gene by the edgeR package.i have 2
transcriptome libraries, one is control and other is experimental.
when i estimate overall dispersion for the dataset.  its gives me warning
message:

Warning message:
In estimateGLMCommonDisp.default(y = y$counts, design = design,  :
  No residual df: setting dispersion to NA

and without dispersion value, i can not find DE genes.now what i do?please
anyone tell me.

these are the commands:

rawdata1 <- read.delim("banana.txt", check.names=FALSE,
stringsAsFactors=FALSE)
> head(rawdata)
              0day ethy
1 contig00001  625  115
2 contig00002  662   88
3 contig00003  466  563
4 contig00004  610   98
5 contig00005   72   21
6 contig00006   82    2
> library(edgeR)
> y <- DGEList(counts=rawdata[,2:3], genes=rawdata[,1:1])
Calculating library sizes from column totals.
> y <- calcNormFactors(y)
> y$samples
     group lib.size norm.factors
0day     1    34450    1.1292933
ethy     1    25198    0.8855096
> plotMDS(y)
Error in cmdscale(as.dist(dd), k = ndim) :
  'k' must be in {1, 2, ..  n - 1}
> Patient <- factor(c(0,1))
> Tissue <- factor(c("N","T"))
> data.frame(Sample=colnames(y),Patient,Tissue)
  Sample Patient Tissue
1   0day       0      N
2   ethy       1      T
> design <- model.matrix(~Patient+Tissue)
> rownames(design) <- colnames(y)
> y <- estimateGLMCommonDisp(y, design, verbose=TRUE)
Warning message:
In estimateGLMCommonDisp.default(y = y$counts, design = design,  :
  No residual df: setting dispersion to NA
> y <- estimateGLMCommonDisp(y)
> y <- estimateGLMTrendedDisp(y)
> y <- estimateGLMTagwiseDisp(y)
Warning message:
In movingAverageByCol(apl[o, ], width = 1000) :
  reducing moving average width to nrow(x)
> fit <- glmFit(y, design)
Error in beta[i, ] <- z %*% X :
  number of items to replace is not a multiple of replacement length
> rawdata1 <- read.delim("banana.txt", check.names=FALSE,
stringsAsFactors=FALSE)
>  library(edgeR)
> y <- DGEList(counts=rawdata[,2:3], genes=rawdata[,1:1])
Calculating library sizes from column totals.
>  y <- calcNormFactors(y)
>  y$samples
     group lib.size norm.factors
0day     1    34450    1.1292933
ethy     1    25198    0.8855096
>  plotMDS(y)
Error in cmdscale(as.dist(dd), k = ndim) :
  'k' must be in {1, 2, ..  n - 1}



please help me.

regards
deepika

	[[alternative HTML version deleted]]