Aggregate data across multiple samples

[jdidion]

It may be useful to generate a single BAM file for parallelization across many samples, rather than one per sample. To do that, we can simply sum the volumes across samples for each interval, and then split on the aggregate data.