To answer my own question, the discrepancy appears to be due to a bug in lastz that causes an incorrect value for L to be printed. I have informed the lastz author of this problem.
Thanks, -Michiel de Hoon RIKEN Yokohama Dear UCSC, To make sure I understand the pairwise genome alignment process, I am trying to regenerate the canFam2-equCab2 alignment. I am getting something close but not quite the same as the alignment available from UCSC. One thing I was wondering about are the parameters used for the lastz runs. According to ftp://hgdownload.cse.ucsc.edu/goldenPath/canFam2/vsEquCab2/README.txt, K=3000 and L=2200 were used (in addition to O=400 E=30 H=2000 M=50). However, the .net.axt.gz files have the following lines in their header: ##aligner=lastz.v1.00.05 H=2000 M=50 ##blastzParms=O=400,E=30,K=3000,L=3000,M=0 suggesting that L=3000 was used instead of L=2200 (I also assume that M=50 was used instead of M=0, since we don't remove repeat regions from these genomes before aligning). Which are the correct lastz parameters to use? Many thanks, Michiel de Hoon RIKEN Omics Science Center, Yokohama _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
