When I opened #17, MANTIS would crash for our GRCh37 bams (reference used and MS bed file where also generated for this release), where chromosome are not prefix chr. It does not do that anymore apparently (although I don't know which change is responsible for this).
However, we have been noticing the following in the log files:
[05/16/18 14:55:23] Loading target MSI loci from BED file ...
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer GCCC.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer CCTC.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer CTG.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer GCT.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer GGT.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer GGA.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer GT.
[05/16/18 14:55:26] Error: Specified locus does not appear to be the starting point for kmer T.
Maybe not surprisingly, we get as many warnings/errors in the log file as there are lines in the bed file.
We have compared the results for 50+ samples with an earlier pipeline run that did not print the warnings - all DIF scores match, so there seems to be no problem here and the bed file should be fine.
I suppose there are still some rough edges in the handling of chromosome names.
This seems to be related to #17 and #22.
When I opened #17, MANTIS would crash for our GRCh37 bams (reference used and MS bed file where also generated for this release), where chromosome are not prefix
chr. It does not do that anymore apparently (although I don't know which change is responsible for this).However, we have been noticing the following in the log files:
Maybe not surprisingly, we get as many warnings/errors in the log file as there are lines in the bed file.
We have compared the results for 50+ samples with an earlier pipeline run that did not print the warnings - all DIF scores match, so there seems to be no problem here and the bed file should be fine.
I suppose there are still some rough edges in the handling of chromosome names.