#83 Options to control A-stat threshold

[Jason Rafe Miller]

The unitig module optionally takes a genome size. Setting this parameter affects the threshold for unique vs repetitive unitigs. When the genome is set artificially small, then the given number of reads are expected to pile up, so high-coverage unitigs will not be marked repetitive. (This from G.S.:) We employ this trick, for instance, in metagenomics assemblies to increase use of the high-coverage unitigs. One tries to set the genome size so that almost all large (those over 10Kbp) contigs would have an astat that causes them to be treated as unique. There is already code with logic like that in both the old unitigger and in BOG, but it is tuned toward looking at the median value of large unitigs. This code could take an option to set the arrival rate to the maximum for large contigs instead.