Entering edit mode
I have data composed of 3 different controls and 3 different treatments each in triplicates. The data is represented below for easier understanding:
sample condition condition_by_groups
controlA_1 untreated untreated_A
controlA_2 untreated untreated_A
controlA_3 untreated untreated_A
controlB_1 untreated untreated_B
controlB_2 untreated untreated_B
controlB_3 untreated untreated_B
controlC_1 untreated untreated_C
controlC_2 untreated untreated_C
controlC_3 untreated untreated_C
treatedA_1 treated treated_A
treatedA_2 treated treated_A
treatedA_3 treated treated_A
treatedB_1 treated treated_B
treatedB_2 treated treated_B
treatedB_3 treated treated_B
treatedC_1 treated treated_C
treatedC_2 treated treated_C
treatedC_3 treated treated_C
A, B, C represents the different controls or treatments and the numbers 1, 2 and 3 the biological repeats.
I would like to analyze the difference in gene expression in the following way:
- treated vs untreated (disregarding the fact that controls and treatments are different)
- pairwise by groups (e.g. treated_A vs untreated_A; treated_B vs untreated_B, etc.)
I have setup the design like this:
coldata <- data.frame(
sample = c( "controlA_1", "controlA_2", "controlA_3", "controlB_1", "controlB_2", "controlB_3", "controlC_1", "controlC_2", "controlC_3",
"treatedA_1", "treatedA_2", "treatedA_3", "treatedB_1", "treatedB_2", "treatedB_3", "treatedC_1", "treatedC_2", "treatedC_3" ),
condition = c( "untreated", "untreated", "untreated", "untreated", "untreated", "untreated", "untreated", "untreated", "untreated",
"treated", "treated", "treated", "treated", "treated", "treated", "treated", "treated", "treated" ),
condition_by_groups = c( "untreated_A", "untreated_A", "untreated_A", "untreated_B", "untreated_B", "untreated_A", "untreated_C", "untreated_C", "untreated_C",
"treated_A", "treated_A", "treated_A", "treated_B", "treated_B", "treated_B", "treated_C", "treated_C", "treated_C" ),
row.names = "sample" )
coldata$condition <- as.factor(coldata$condition)
coldata$condition_by_groups <- as.factor(coldata$condition_by_groups)
#construct DESeqDataSet1
dds <- DESeqDataSetFromMatrix(countData = count_matrix, colData = coldata,
design = ~ condition)
dds <- DESeq(dds)
res0 <- results(dds, contrast=c("condition", "treated", "untreated"))
#construct DESeqDataSet2
dds <- DESeqDataSetFromMatrix(countData = count_matrix, colData = coldata,
design = ~ condition_by_groups)
dds <- DESeq(dds)
res1 <- results(dds, contrast = c('condition_by_groups', 'treated_A', 'untreated_A'))
res2 <- results(dds, contrast = c('condition_by_groups', 'treated_B', 'untreated_B'))
res3 <- results(dds, contrast = c('condition_by_groups', 'treated_C', 'untreated_C'))
Is it appropriate for the analyses I want to perform or something needs to be modified?