Scout: CLI command to view exons seems not working

[northwestwitch]

Check the command:
scout view exons

that returns just one exon:

{'_id': ObjectId('5f633f4f582d613b6c72dc65'),
'build': '37',
'chrom': '19',
'end': 58864804,
'exon_id': '19-58864770-58864804',
'hgnc_id': 5,
'rank': 1,
'start': 58864770,
'strand': -1,
'transcript': 'ENST00000263100'}

This command might be redundant since we already have scout export exons, so check what's the difference between the 2.

[northwestwitch]

Ok now I'm really confused. I've checked on the database and the exon collection of the production db is EMPTY. The one of stage db has only exons from build 38.. 🤔

[northwestwitch]

Update on this one. It might not be a bug, but just us that we didn't load the exons. I'm running the scout load exons command on stage now. Then I'll see if the export works.

If everything is fine I'll run it on prod as well

[northwestwitch]

It was that. Shame on me, but an easy (partial) fix!
After loading the exons, the scout export exons command works, but the scout view exons one doesn't.

l'll just leave the issue open and rename it because it needs fixing, but it's less urgent

[dnil]

Sounds good - check with @moonso otherwise - I still think he was the last one touching that?

[northwestwitch]

Sounds good - check with @moonso otherwise - I still think he was the last one touching that?

He wrote the code to load exons and to export them (with scout export), and both things work. It's the view command that needs a check.

[moonso]

Hi, great that it worked! The view vs export is a bit vague. The first idea was that view where supposed to be used when doing human inspections of stuff in the database, so readable format. The export commands was more for serialising things. Since I guess none of them are used so much one could give it a thought and perhaps merge them.

[northwestwitch]

I think that if we merge them we should still keep the option to query just for the exons of one gene, instead of all exons from all genes.

[henrikstranneheim]

Do we not use the scout export exons to get input to the sambamba -> chanjo -> coverage track

[dnil]

I agree that was one of the ideas, but its unclear to me how or if the flow of information is automated - @moonso again? I guess via cg. It would be the tool uploading runs to chanjo anyway, so why not also pull exon info from scout to upload? But then it would be in trouble if the exon db is empty on prod? Perhaps that part was never implemented, only done a few times manually?

Scout afaik currently only speaks indirectly to chanjo via chanjo-report, and can then only give gene symbols, not new exon definitions.

[dnil]

ok so two parts

• [ ] scout view exons without options (possibly with an empty exon db) does not produce the expected result
• [ ] possibly harmonise view exons and export exons, but from @moonso's answer I think it may not be necessary?