Framefix introducing errors?

[emdyess]

Hi Paul,

I have run framefix on a few bacterial hybrid assemblies (ONT+Illumina) generated with Unicycler or Flye/Racon/Pilon. I am using the Uniref90 database:

/path/to/framefix /path/to/assembly.fasta 1e-25 /path/to/uniref90.dmnd

I have a "Gold Standard" PacBio reference sequence that I am comparing the ONT+Illumina hybrid assemblies to, and when I run NucDiff against my assembly without framefix and then my assembly with framefix, it appears that errors are introduced. For example (Pseudomonas aeruginosa):

Without framefix:
Insertions: 73
Deletions: 28
Substitutions: 5

With framefix
Insertions: 99
Deletions: 57
Substitutions: 5

Do you have any idea why this might be? Should I try a different database?

Thanks!

[nodrogluap]

Thanks for trying it out. To figure out what is going on, could you post the output_prefix.diamond_nr.txt file? This reports the database sequence matches that were used to adjust each frameshift.