In TeamMacLean/atacr: Analysing Capture Seq Count Data

The atacr Object

When make_counts() is run, an atacr object is returned. This is a simple, somewhat informal object based on the R list type. It is basically an R list with the following members:

1. treatments - a character vector of treatment names
2. samples - a character vector of sample names
3. bam_files - a character vector of paths for the used BAM files
4. bait_regions - a GenomicRanges::Granges object describing the bait window regions
5. bait_windows - a RangedSummarizedExperiment object containing the counts in the windows in bait_regions
6. non_bait_windows - a RangedSummarizedExperiment object containing the counts in the windows in the regoions outside bait_regions
7. whole_genome - the union of bait_windows and non_bait_windows
8. dataframe - an optional member and the result of calling as.data.frame() on the atacr object

Column Order

The RangedSummarizedExperiment objects carry the count data. They are organised as a matrix with rows representing windows and columns different samples. Their order is conserved and is the same as that in the treatments, samples and bam_files.

The 'which' argument

Many of the functions allow you to state which member of the atacr list (really a RangedSummarizedExperiment) you wish to apply the function to with the which argument, e.g

plot_counts(counts, which = "bait_windows", log10 = FALSE)

Adding members to the atacr object

In this way you can use functions that return RangedSummarizedExperiments to become new members in the list and work on them as with the built in ones, this is especially useful for normalisations.

counts$by_sample <- library_size_normalisation(counts, 
                                             by_treatment = TRUE)

plot_counts(counts, which = "by_sample", log10 = FALSE)

TeamMacLean/atacr documentation built on May 9, 2019, 4:24 p.m.