DESeq without replicates

[gfmgfm]

Hello all,

Does DESeq use replicate info in estimateSizeFactors?

For RNA seq without replicates - will it be OK to normalize the data using:
estimateSizeFactors
(and then dividing each column by the size factors?)

Also if I have several conditions, and have a replicate only for one of them - is it OK to use estimateSizeFactors?

Thanks.

[dpryan]

Does DESeq use replicate info in estimateSizeFactors?

No, at least not in the way you probably mean (I'm reading into the remainder of your post a bit). estimateSizeFactors doesn't measure dispersion, but rather is used to account for, for example, the different depths of sequencing of the various samples.

For RNA seq without replicates - will it be OK to normalize the data using:
estimateSizeFactors
(and then dividing each column by the size factors?)

Also if I have several conditions, and have a replicate only for one of them - is it OK to use estimateSizeFactors?

You'll need to use that with most experimental setups. Have a read through the DESeq vignette, it'll help flush out what estimateSizeFactors is used for and what to do in cases where replicates are lacking (if you totally lack them, the answer is "you should really sequence some replicates!").

[gfmgfm]

Thanks a lot for the reply.
Yes I meant using the SizeFactors only for normalization due to different coverage between the samples.
I know one must have replicate - but still don't have them all