Test cases:
    o NM_004038.3
    o gapOver
    o many
    o misc
    o noDropCntBug
    o overlap
    o pseudoRetro
    o repMatch
    o haplotype - hg18
      o hapMappings.psl - blat alignment of hap chroms to chroms
        omitted chr6_qbl_hap2 aligments because it's so larger
      o examples:
        - NM_052923.1 off edge of chr6_cox_hap1 region on chr6
        - NM_000976.2 - aligns to chr6, chr6_qbl_hap2; best align is chr9
        - NM_001037638.1 - chr6, chr6_qbl_hap2, chr6_cox_hap1
        - NM_207331.2 - dups on chr5, chr5_h2_bhap1;  should keep 3 each
        - NM_001033523.1 - multiple alignments on chr5, chr5_h2_hap1, best is chr6
        - NM_002116.5 - chr6, chr6_qbl_hap2, chr6_cox_hap1; should keep all

    o tandemGeneDup - hg18 - tandem duplicated genes were blat maybe joining two different
        copies in an alignment
        - BC029540
