Calling Diploid Genotypes with MPG

The MPG ("Most Probable Genotype") genotype caller takes as its input 
a set of sequence reads aligned to an orthologous reference sequence. 
These sequence reads are assumed to be (1) correctly aligned to the 
reference, and (2) derived from one or two copies of a modified 
reference sequence, as would be the case in re-sequencing reads from 
non-repetitive regions of a diploid individual. In addition, if base 
quality estimates are provided, these are assumed to be accurate 
reflections of the probability that the read sequence actually 
represents the sample. The MPG algorithm is implemented in a perl script 
called "bam2mpg", which reads BAM formatted files, and outputs predicted 
genotypes and scores at all regions of the genome covered by at least 
one read. 
