amrfinder correct for read counts

[andrewdavidsmith]

• smithlab_cpp: updating submodule
• amrfinder: added an option to turn off the correction to likelihood for the number of reads which is applied to the likelihood for the two alleles model
• EpireadStats.hpp: changing the EpireadStats class to struct and allowing default values for default constructor. Also adding an instance variable to indicate whether log likelihood for allel pair model should correct for number of reads
• EpireadStats.cpp: adding a variable to control the use of correction to the allele pair model log likelihood based on number of reads
• amrfinder.cpp: Using the default constructor for EpireadStats
• amrtester: using an approach more similar to amrfinder which uses the EpireadStats class instance and allows a correction for read count for the allel pair model log likelihood
• amrtester: more changes
• md5sum: update for the empty amr file output by amrfinder currently