Hi,

I'm trying to run Deseq2 analysis (R-4.1.1, Deseq2 1.31.16) on RNA Seq data. We did targeted RNA Seq on 10 genes for expression data (only 10 amplicons). It's the commercial targeted RNAseq Assay from thermofisher scientific (S5 sequencing). I have about 200 samples: 10 controls and 190 cases; and 5 housekeeping genes and 5 genes of interest. I would like to get normalized counts and log2 fold change to make a comparison with RqPCR data on the same genes of interest. For my qPCR data, I calculated ΔCt (betwin housekeeping genes and genes of interest) for cases and controls, then ΔΔCt for cases and finally the 2^-(ΔΔCt).

I would like to get the equivalent for my RNAseq data. I tryied Deseq2 with tutorials but I only obtained log2fold change by genes... I don't know how to obtain fold changes by cases.

Important information: I'm not a bioinformatician :)

Used code for the first and ineffective analysis:

*library("DESeq2")
library(ggplot2)
setwd("E:/R-4.1.1/library/DESeq2")
countData<-read.csv('DataNGS.csv',header=TRUE,sep=",")
metaData<-read.csv('metadata3.csv',header=TRUE,sep=",")
#Construct DESEQDataSet Object
dds<-DESeqDataSetFromMatrix(countData=countData,colData =metaData,design=~dex,tidy=TRUE)
#Run DESEQ function
dds<-DESeq(dds)
#Check result table
res<-results(dds)
head(results(dds,tidy=TRUE))
summary(res)
#Sort summary list with p-value
res<-res[order(res$padj),]
head(res)
#plotCounts
par(mfrow=c(2,3)) 
plotCounts(dds,gene="0-EIF2B1.E7E8.110",intgroup="dex")
plotCounts(dds,gene="1-FBXW2.E3E4.110",intgroup="dex")
plotCounts(dds,gene="2-PSMB2.E1E2.110",intgroup="dex")
plotCounts(dds,gene="3-PUM1.E21E22.110",intgroup="dex")
plotCounts(dds,gene="4-TRIM27.E2E3.110",intgroup="dex")
plotCounts(dds,gene="5-MECOM.E6E7",intgroup="dex")
plotCounts(dds,gene="6-WT1.E7E8",intgroup="dex")
plotCounts(dds,gene="7-BAALC.E2E3",intgroup="dex")
plotCounts(dds,gene="8-MYC.ENCTRL.E2E3",intgroup="dex")
plotCounts(dds,gene="9-SMC1A.E21E22",intgroup="dex")*

Thanks! Claire

The ΔΔCt that you calculated from your qPCR data are normalized log fold changes, and are equivalent to what you get from DESeq2. Converting by doing 2^-(ΔΔCt) linearizes the fold change, which is often helpful for the log averse amongst us, but in general I think it is probably less useful than one might think.