Invalid aminoacid change detected due to GTF file phase ignored

[znikasz]

Describe the bug
Using hg38_ensembl.ser, version 95 jannovar annotates change

chr1 939278 C > T

in transcript ENST00000455979 as p.(R2C). This is incorrect, the correct change is A1V.

In my opinion Jannovar ignores the frame shift (phase) data in GTF file, which is 2 in the case of ENST00000455979:

1       havana  CDS     939275  939460  .       +       2       gene_id "ENSG00000187634"; gene_version "11"; transcript_id "ENST00000455979"; transcript_version "1"; exon_number "1"; gene_name "SAMD11"; gene_source "ensembl_havana"; gene_biotype "protein_coding"; transcript_name "SAMD11-205"; transcript_source "havana"; transcript_biotype "protein_coding"; protein_id "ENSP00000412228"; protein_version "1"; tag "cds_start_NF"; tag "mRNA_start_NF"; transcript_support_level "2";

To Reproduce
Steps to reproduce the behavior:

1. Annotate chr1 939278 C > T and check it's protein change in ENST00000455979 .

Expected behavior
Aminoacid change should be A1V

Additional context
Please check if you don't ignore phase in CDS entries.

[holtgrewe]

Confirmed, phase is indeed ignored.

[holtgrewe]

We need to ask ENSEMBL of the meaning of frame information for CDS on reverse strand for proper implementation.