Question: Deseq2 Variance Stabalizing Transformation -- loss zeros to standard integer
0
gravatar for chicoin1
8 weeks ago by
chicoin10
chicoin10 wrote:

Hello,

I used Deseq2's Variance Stabalizing Transformation to normalize 16S amplicon otu table. Unlike other users, I have no negative values in my final dataset. Further, I have no 0s in the dataset and, after comparing the transformed otu table to the raw data, I noticed that all of the otus that were 0 in the raw data now have a consistent integer (1.015063211).

Is there underlying code in the VST function that adds a constant integer to all samples? Would it be appropriate to convert all cells with this integer to 0? I'm interested in determining which otus are shared among my treatments, but right now I can't do this because I have no 0s in my dataframe (e.g. can't compare presence/absence across the samples).

Code for transformation: bact.p.deseq=phyloseqtodeseq2(bact.p, ~X.SampleID) vsd<-varianceStabilizingTransformation(bact.p.deseq)

Thanks for your help!

ADD COMMENTlink modified 8 weeks ago by Michael Love23k • written 8 weeks ago by chicoin10
Answer: Deseq2 Variance Stabalizing Transformation -- loss zeros to standard integer
0
gravatar for Michael Love
8 weeks ago by
Michael Love23k
United States
Michael Love23k wrote:

Take a look at our RNAseq workflow, you can see what’s happening across vst(), rlog() and the shifted log. The above behavior is expected.

“I'm interested in determining which otus are shared among my treatments”

I don’t think the VST helps you answer this question. Maybe another package specific to 16S would be more appropriate.

ADD COMMENTlink written 8 weeks ago by Michael Love23k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 120 users visited in the last hour