Evaluation modes and evaluation flags

Evaluation flags

Many import QC steps need to be performed on every genome (such as confirming the nucleotide sequence is not already present in the database under a separate name). However, a MySQL database may store data for diverse types of genomes, and some evaluations are dependent on factors such as the annotation status, the authorship, or the data source. As a result, some QC steps can be turned on (yes) and off (no) depending on the type of genome being imported.

check_replace

Should unexpected genome replacements be reported?

import_locus_tag

Should CDS feature locus_tags be imported?

check_locus_tag

Should the structure of CDS feature locus_tags be evaluated?

check_description_tally

Should the number of descriptions be evaluated?

check_description_field

Should CDS descriptions in unexpected fields be reported?

check_description

Should unexpected CDS descriptions be reported?

check_trna

Should tRNA features be evaluated?

check_id_typo

Should genome ID typos be reported?

check_host_typo

Should host typos be reported?

check_author

Should unexpected authors be reported?

check_gene

Should the CDS ‘gene’ qualifier be evaluated?

check_seq

Should the nucleotide sequence be evaluated?

check_coords

Should feature duplication be evaluated?

Evaluations modes

In order to manage which evaluations are implemented, the evaluation mode is specified for each ticket:

draft

For new, automatically-generated ‘draft’ annotations, since they are likely to have several types of automatically-generated errors that can be ignored.

final

For new, manually reviewed ‘final’ annotations that are expected to have very few errors.

auto

For manually reviewed ‘final’ annotations that are automatically retrieved from GenBank. Some errors are ignored since they are already publicly available.

misc

For genome annotations created from an external source. Since it is not always known how they have been annotated, and since any errors may not be able to be changed, certain types of errors are ignored.

custom

For manual selection of the specific evaluation flags that should be performed if none of the other four preset run modes are appropriate.