My dataset: 3 different cell types, with 3 replicates each

Celltype is the factor and "Alveolarmacrophages" "Interstitialmacrophages" "Tcells" are the factor levels.

I created the DESeq object:

>dds <- DESeqDataSetFromMatrix(countData = only_counts,
                              colData = sample_info,
                              design= ~ Cell_type)

and calculated the results from the comparison of one factor level against the others:

>res <- results(dds, alpha=0.05)

So the DESeqResults object has three elements:

> resultsNames(dds) 
[1] "Intercept"                                                  "Cell_type_Interstitial_macrophages_vs_Alveolar_macrophages"
[3] "Cell_type_T_cells_vs_Alveolar_macrophages"   

Now I want to access ONE of these elements for further analysis. When I type 'res' I only see one of them

> res
log2 fold change (MLE): Cell type T cells vs Alveolar macrophages 
Wald test p-value: Cell type T cells vs Alveolar macrophages 
DataFrame with 1978 rows and 6 columns

So I would also like to access the other one, "Cell_type_Interstitial_macrophages_vs_Alveolar_macrophages". I have tried:

> res["Cell_type_Interstitial_macrophages_vs_Alveolar_macrophages"]
Error: subscript contains invalid names
>res[[2]]   #super large vector
> res_am <- results(dds, coef=2)
Error in results(dds, coef = 2) : unused argument (coef = 2).  #I find this error weird since the same is used in the DESeq vignette in the LFC example

When I do

>res_im_vs_am <- results(dds, name="Cell_type_Interstitial_macrophages_vs_Alveolar_macrophages")

this is not really accessing the element from 'res', it´s recalculating the results, all the information that I included when I did res<-results(dds,alpha=0.05) is lost; e.g. that alpha=0.05, or any other customization I could have added.

Is there any way to directly access the elements that resultsNames(dds) shows?

Right now I only have 3 levels, but later I will add others, so it doesn't make sense to do contrast by contrast.

Is there any way to do all the possible comparisons in a single line of code and store the resulting data frames in e.g. a list?

"all the information that I included when I did res<-results(dds,alpha=0.05) is lost;"

Why aren't you providing those arguments to your results() call?

Think of results() as an extractor, it doesn't modify dds which is the dataset object. This is why we have the paradigm:

dds <- modify(dds)
dds <- modify(dds)
...
res <- extract(dds)

We change the variable name in the documentation and vignette to indicate that the last function is not returning the dataset.

So if you want a results table for interstitial vs alveolar with target alpha of 0.05 you should call results() while specifying those as arguments.

"Is there any way to do all the possible comparisons in a single line of code and store the resulting data frames in e.g. a list?"

No, but you can easily call results() multiple times.