Entering edit mode
Hi,
has anyone used a normalization strategy, called remove unwanted variation (RUV) (Risso et al. 2014, Nature), in DESeq analyses?
6
Entering edit mode
The methods RUVg and RUVs should be easy to incorporate with DESeq2, by adding these factors to the colData(dds) and including them in the design. The RUVr method requires an edgeR output as input. A similar approach to estimating these factors is svaseq, and we have a demonstration of how this would look in the workflow.
0
Entering edit mode
Thanks for your quick reply! However, I am afraid that I might have not formulated my question correctly as I would like to know whether it is possible to use RUVg corrected counts as input in DESeq2 analysis. The reason why I would like to use the RUV normalization is that I have few samples that have highly biased expression profiles. These few samples are lacking some extremely highly expressed genes that are present in the rest of the samples.This leads to a situation where the rest of the genes accumulate more reads in these few samples. I am afraid that this will somehow mess with the size factor estimation.
Similar Posts
Loading Similar Posts
Traffic: 643 users visited in the last hour
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by the
version 2.3.6
version 2.3.6
No you wouldn't use the correct counts, but just the factors of unwanted variation. Then, this will be accounted for in the model for genes which have the biased expression you see.
Some highly expressed or biased genes shouldn't affect size factor calculation because it uses the median over all genes. Or you can estimate size factors however you like and import them with:
before running DESeq().