Issue:
Sometimes when annotating a huge MAF, some variants do not get a protein change or the annotation times out. Re-running those failed records produces the amino acid change.
Solution:
Would be nice to add a flag to the annotate subcommand or let the default behaviour be such that if all the records are not successfully annotated on the first attempt, then the script will continue running the annotator on the remaining unannotated records until no new annotated records are produced in further attempts.
Logic:
Let the annotation run normally the in the first attempt.
Subset the variants which failed annotation i.e, the Variant_Classification is not in ["Silent", "Intron", "3'UTR", "5'UTR", "3'Flank", "5'Flank", "IGR"] and the HGVSp_Short is empty or Annotation_Status is FAILED. Basically, split the MAF to annotated and unannotated parts.
Re-run the annotation of the unannotated MAF and merge the annotated records to the annotated MAF.
Repeat until all the variants are annotated or no new annotated records are produced in the further attempts.
Issue: Sometimes when annotating a huge MAF, some variants do not get a protein change or the annotation times out. Re-running those failed records produces the amino acid change.
Solution: Would be nice to add a flag to the
annotatesubcommand or let the default behaviour be such that if all the records are not successfully annotated on the first attempt, then the script will continue running the annotator on the remaining unannotated records until no new annotated records are produced in further attempts.Logic:
Variant_Classificationis not in["Silent", "Intron", "3'UTR", "5'UTR", "3'Flank", "5'Flank", "IGR"]and theHGVSp_ShortisemptyorAnnotation_StatusisFAILED. Basically, split the MAF to annotated and unannotated parts.Refer to the python implementation here for more details - https://github.com/cBioPortal/datahub-study-curation-tools/tree/master/GN-annotation-wrapper