**1.2k**wrote:

Dear all,

Our RNA-seq data has two samples, each of which has three biological replicates. I have two edgeR scripts as follows. I found that they produced different results. My question is: which model (classical linear or GLM) to choose in edgeR analysis?

I wish biostars experts on R analysis could provide me some comments or suggestions. Thank you very much .

**- R script 1**

```
d <- read.delim("input", sep = ' ', header=F, row.names=1)
colnames(d) <- c("w1", "w2", "w3", "m1", "m2", "m3")
g <- factor(c(1,1,1,2,2,2))
dge <- DGEList(counts=d,group=g)
dge <- calcNormFactors(dge)
design <- model.matrix(~g)
dge <- estimateDisp(dge, design)
fit <- glmQLFit(dge, design)
qlf <- glmQLFTest(fit, coef=2)
topTags(qlf, n=100000)
```

**R script 2**

```
d <- read.delim("input", sep = ' ', header=F, row.names=1)
colnames(d) <- c("w1", "w2", "w3", "m1", "m2", "m3")
g <- factor(c(1,1,1,2,2,2))
dge <- DGEList(counts=d,group=g)
dge <- estimateCommonDisp(dge)
dge <- estimateTagwiseDisp(dge)
et <- exactTest(dge)
topTags(et, n=100000)
```