Variance assembled but not called

[ssjunnebo]

From a recent path-help ticket:

I have a strange occurrence in ariba, where it assembles the gene correctly, but does not flag the variance mutation (even though the assembled gene I get back from ariba actually has that mutation) because it predicts a frameshift, but assembles the genes correctly without indicating a frameshift.

The files are in /lustre/scratch118/infgen/team15/eh10/Klebsiella_Addenbrookes/Enterob_intrinsic

The variance files are in var_database/; and e.g. ampD from 9776_8#36 (ariba_9776_8#36_var_out/) has the M101K mutation, also in the sequence ariba gives as assembled output, but this does not get recognised in the report file – other mutations (e.g. F63Y) get recognised correctly however for other strains, so there does not seem to be a problem with the database as such. Do you have any suggestions on why this is happening? It predicts a frameshift at position 70, but when I look at the assembled gene it translates fine without frameshift? My references don't have a frameshift either that could cause the problem?