Hi all,
I am designing a new experiment to sequence transcriptomes from wasps. I have 3 treatment "environments" and 4 replicate populations in each (kept separate throughout), for a total of 12 separate rearing cages.
I want to compare expression among treatment "environments", but of course the replicate populations should be taken into account. They are separate pools of individuals that have evolved separately. The problem is that my experimental wasps are not very cooperative, and I simply don't have 3 individuals from some of my replicate populations. So, I am trying to decide how to go about sequencing them.
Can I assume similar variance across populations, and so get a good estimate of expression from non-replicated populations? Is this a terrible idea because they are coming from different genetic backgrounds?
For some of my replicate populations, I have plenty of individuals. For others, just one. So, I am wondering how useful it will be to sequence those single individuals. As an alternative, I can use my limited resources differently, and just sequence more individuals from fewer replicate populations. This will tell me less about my experiment in general, but more about that population.....
According to this thread, I can analyze using edgeR if I just have good replication in some groups: http://seqanswers.com/forums/showthread.php?t=46068
Does anyone have experience with this? In EdgeR or other programs?
For comparison/context, in previous sequencing from these populations, the wasps were more cooperative and I used the following design to analyze differential expression in DESeq2 (35 libraries in total):
sel.line <- factor(c(rep("GroupB",11),rep("GroupD",12), rep ("GroupE", 12)))
block <- factor(c(rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 2),
rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 3),
rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 3)))
ExpDesign3 <- data.frame(
row.names = colnames(countdata3),
sel.line = sel.line,
block = block)
CDS3 <- DESeqDataSetFromMatrix(countData = countdata3, colData=ExpDesign3, design= ~ block + sel.line)
"
In this case, block refers to replicate cages within treatment, and sel.line is the level of "interest" in my study.
Many thanks,
Alice
I am designing a new experiment to sequence transcriptomes from wasps. I have 3 treatment "environments" and 4 replicate populations in each (kept separate throughout), for a total of 12 separate rearing cages.
I want to compare expression among treatment "environments", but of course the replicate populations should be taken into account. They are separate pools of individuals that have evolved separately. The problem is that my experimental wasps are not very cooperative, and I simply don't have 3 individuals from some of my replicate populations. So, I am trying to decide how to go about sequencing them.
Can I assume similar variance across populations, and so get a good estimate of expression from non-replicated populations? Is this a terrible idea because they are coming from different genetic backgrounds?
For some of my replicate populations, I have plenty of individuals. For others, just one. So, I am wondering how useful it will be to sequence those single individuals. As an alternative, I can use my limited resources differently, and just sequence more individuals from fewer replicate populations. This will tell me less about my experiment in general, but more about that population.....
According to this thread, I can analyze using edgeR if I just have good replication in some groups: http://seqanswers.com/forums/showthread.php?t=46068
Does anyone have experience with this? In EdgeR or other programs?
For comparison/context, in previous sequencing from these populations, the wasps were more cooperative and I used the following design to analyze differential expression in DESeq2 (35 libraries in total):
sel.line <- factor(c(rep("GroupB",11),rep("GroupD",12), rep ("GroupE", 12)))
block <- factor(c(rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 2),
rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 3),
rep("1", 3), rep("2", 3), rep("3", 3), rep("4", 3)))
ExpDesign3 <- data.frame(
row.names = colnames(countdata3),
sel.line = sel.line,
block = block)
CDS3 <- DESeqDataSetFromMatrix(countData = countdata3, colData=ExpDesign3, design= ~ block + sel.line)
"
In this case, block refers to replicate cages within treatment, and sel.line is the level of "interest" in my study.
Many thanks,
Alice