I would like to know how to apply lfcShrink to different contrasts where I have 4 groups in my model (LOCHYcontrol, LARUcontrol, LOCHYpolyIC and LARU polyIC).

The contrasts I am interested in are;

1. LOCHYcontrol vs LARUcontrol
2. LOCHYpolyIC vs LARUpolyIC
3. LOCHYcontrol vs LOCHYpolyIC
4. LARUcontrol vs LARUpolyIC

When I run the model and look at the resultsNames, I am not sure how to then apply lfcShrink to the contrasts I am interested in as only one of these appears in the resultsNames (site_treat_LOCHYpolyIC_vs_LOCHYcontrol).

Is there a way to access all possible contrasts with lfcSkrink?

### create agglomerated variable ###
coldata$site_treat <- as.factor(paste(coldata$site, coldata$treatment, sep=""))
coldata$site_treat <- relevel(coldata$site_treat, "LOCHYcontrol","LARUcontrol","LOCHYpolyIC","LARUpolyIC")
coldata

### using agglomerated variable to model counts, so 4 levels ###
dds_new <- DESeqDataSetFromMatrix(countData = counts,
                                  colData = coldata,
                                  design = ~ site_treat)

dds_new

### run DESeq analysis ###
dds_new <- DESeq(dds_new)

resultsNames(dds_new)

## [1] "Intercept"        
## [2] "site_treat_LARUcontrol_vs_LOCHYcontrol"
## [3] "site_treat_LARUpolyIC_vs_LOCHYcontrol"  
## [4] "site_treat_LOCHYpolyIC_vs_LOCHYcontrol"

### to run lfc shrinkage after the fact ###
resLFC <- lfcShrink(dds_new, coef=4, type="apeglm")
resLFC

You have to relevel the group and then rerun the Wald test in order to shrink all contrasts, here is a code suggestion: https://www.biostars.org/p/448959/#484944

Or just use ashr which takes contrasts out of the box.