Sequences are different in Reference sequence track and Annotations track

[shashankbrgowda]

Reported by Toby

The genome sequence displayed in the "Reference sequence" track is completely different to that displayed in the "Annotations" track (or more accurately out of sync with each by a few bases).

This also seem to result in the splice site checker in the "transcript details" window showing lots of non-canonical splice sites. In the case below (ENST00000661030) I know all the annotated splice sites are actually canonical but are not being displayed as such:

However, when I refresh the page this error resolves itself. The genome sequence in both tracks is now the same (although the "horizontal flip" function works slightly differently in each), and the splice sites for ENST00000661030 appear correctly:

It seems like if you start at these coordinates: 1:78,938,247..78,938,318[rev] and then start navigating around/altering the co-ords a bit (i.e. by a combination of zooming in and out; flipping the strand etc) you can normally replicate the issue fairly often.....

[garrettjstevens]

@shashankbrgowda and I reproduced this by zooming all the way in before opening the tracks in a fresh session.

[dariober]

This issue was due to incorrect merging of overlapping regions and it should be resolved here.

There was also a related issue with fetching the end of a reference sequence: Load packages/apollo-cli/test_data/tiny.fasta.gff3, select the "reference sequence track" (annotation track is fine) and go to the end of ctgA, you get:

This is because method getFeatures in ApolloSequenceAdapter tries to access a region extending beyond the end of the contig. For now this is fixed here by simply resetting the stop coordinate if the sequence is shorter than declared. It would be better to find out why it happens though...