It shouldn't make a difference, but this way we avoid having to output string representations of floats -- since the numerator (# mutations in sufficiently-covered positions) and denominator (# sufficiently-covered positions) must both be integers. I mean, in order to compute the indices the reader of this file will eventually have to divide these quantities, so there is a tradeoff between ease of use (no need for user to do division) and pathological avoidance of floating-point stuff.
... Probably safe to keep this as is, but may be worth thinking about.
It shouldn't make a difference, but this way we avoid having to output string representations of floats -- since the numerator (# mutations in sufficiently-covered positions) and denominator (# sufficiently-covered positions) must both be integers. I mean, in order to compute the indices the reader of this file will eventually have to divide these quantities, so there is a tradeoff between ease of use (no need for user to do division) and pathological avoidance of floating-point stuff.
... Probably safe to keep this as is, but may be worth thinking about.