Question about multi-factor Deseq analysis
Entering edit mode
FaisalH • 0
Last seen 3.5 years ago

Hi there!

I would like to use DESeq2 for microbiome data to investigate gene abundance on different groups.

To make it simple, this is information about my samples. I have 2 main groups of patient. 1- with disease 2-matched control to each sample. I have 2 samples for each patient at 2 time-points (TP). 

I'm trying to run deseq2 to investigate the gene differential expression between two time-points of groups with disease and control. Then between same time-point between disease and control group. This picture should make the analysis strategy more clearer 

I split my data to 4 groups. 1- Disease group: First and last TP, 2- Control groups: First and last TP, 3- First TP: Disease and control, and 4- Last TP: Disease and control groups.

I have 2 columns in my metadata/factor table. First coldata Group::First and Last. Second column Condition and each 2 samples from one paitient paired in one number. For example, each sample at Early collection and Late collection from one patient given one number.

This is the design I used for my model

AllData$Group <- factor(AllData$Group)
AllData$Condition <- factor(AllData$Condition)

dds <- DESeqDataSetFromMatrix(countData = GenesCount,
                              colData = AllData,
                              design= ~ Group + Condition)


design(dds) <- formula(~ Group + Condition)

dds <- DESeq(dds, betaPrior=FALSE)


However, when I run the first group pf data of Disease group for both Early and Late collection samples I got this for resultsNames(dds)

 [1] "Intercept"              "Group_Disease_Late_vs_Disease_Early" "Condition_2_vs_1"       "Condition_3_vs_1"       "Condition_4_vs_1"      
 [6] "Condition_5_vs_1"       "Condition_6_vs_1"       "Condition_7_vs_1"       "Condition_8_vs_1"       "Condition_9_vs_1"      
[11] "Condition_10_vs_1"      "Condition_11_vs_1"      "Condition_12_vs_1"      "Condition_13_vs_1"      "Condition_14_vs_1"     
[16] "Condition_15_vs_1"     


I believe this analysis compare every paired samples to the first paired number 1 ! which look strange for me.

I am wondering is my model for this experiment is correct ?  any advice to improve the analysis methodology?

Thank you!

DESeq deseq2 microbiome gut microbiome • 468 views
Entering edit mode
Last seen 1 day ago
United States

Your description of the design sounds exactly like something explained in the vignette:

Can you take a look at this section first?


Login before adding your answer.

Traffic: 643 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6