"Best common allele pair" can miss possibilities where one of the best common alleles is in different positions of the pairs #11
Description
Noticed this while doing round 1 of Thunderdome. We can run into situations where the "best common allele pair" is reported sub-optimally. For example, if you see alleles
B*57:01:01 - B*57:02:03B*57:02:04 - B*57:03
then the reported "best common allele pair" is B*57 - B*57, which isn't wrong, but the best possible result here is B*57 - B*57:02.
Update the algorithm for identifying this pair to catch these situations.