Is it possible to create a new dba object based on a subset of the samples in a previous call to dba.count, or dba.analyze?

Basically, I would like analyze subsets of samples as seperate analyses, without recounting.  I'm analyzing atac-seq data, which is paired end and >50 million reads per sample for 20-30 samples.  I've been using the summarize overlaps method to count (bSummarizeOverlaps=TRUE), which works great but takes a lot of time even on a server.     

A second and related question:  is there a way to pull out a summarizedExperiment from dba.report without limiting to  a specific contrast (or with all contrasts)?  I'm trying to get all of my samples into a summarizedExp object, using the counts and peaks data in the dba object.

Thanks for the great package!

Ashley

Hi Ashley-

The dba() function will let you do both of these things.

Using the supplied tamoxifen resistance dataset as an example:

> data(tamoxifen_counts)
> tamoxifen

You can get a subset of these as a new DBA object:

> responsive <- dba(tamoxifen, mask=tamoxifen$masks$Responsive)
> responsive

You can also use dba() to get a SummarizedExperiment version of the DBA object:

> resp.sumEx <- dba(responsive,bSummarizedExperiment=TRUE)
> resp.sumEx

You could do both of these in one step as well:

> resp.sumEx <- dba(tamoxifen, mask=tamoxifen$masks$Responsive, bSummarizedExperiment=TRUE)

Cheers-

Rory

Thanks Rory!