RPKM - Mapped vs. mappable reads

[cmorganl]

Include a flag that allows the user to choose whether they want the RPKM calculations to be normalized by either the number of reads in the SAM file (default) or whether they should be normalized by the number of reads that actually mapped to a sequence. This will need to be parsed from the asterisk in filed 3 to effect the addition of coverage.numreads in rpkm.c++.

Useful for analyses comparing environments where you would want to divide by mappable (all) reads. However, if someone wants to compare bins, sequences, etc. within the same environment, normalizing by number of mapped reads would be best to ascertain the proportion of reads originating from the sequence or population.