Table 1 Goals and methods for the validator.
From: Validation and quality assurance for genome browser database exports
Completeness: Verify that all data has been exported. | Count base features in original database, ensure the count of features in output is an exact match. |
Correctness: Ensure that the content exported accurately represents the original copy. | Manual checking of a subset of features is a pragmatic way to ensure correctness, at least initially. |
Accuracy: Check the placement of annotations relative to references to ensure there are no off-by-one errors. | Select a set of features to use as reference. Calculate distance to other features in both FPD database and GFF3 file. Make sure that features in tracks that are directly related (e.g. exonerate results are derived from BLAST results) correspond at a high rate using a feature comparator such as ParsEval (part of the genometools package which is available at http://genometools.org/). |
Standards Compliance: Verify output conforms to GFF3 standard http://www.sequenceontology.org/gff3.shtml | Use gff3validator (also part of the genometools package) to ensure compliance with the GFF3 standard. |