Fix: Use a consistent country name across usvi samples both submitted to genbank and not yet #41
Conversation
ingest/config/annotations.tsv
Outdated
| KY328289 date 2016-05-15 | ||
| MW165884 country USVI |
There was a problem hiding this comment.
not a request for changes
I was expecting a change in the (local) geolocation rules, but this works too!
There was a problem hiding this comment.
That was my expectation too, since this is overwriting the geolocation rules from ncov-ingest, which include
$ grep -i 'virgin islands' source-data/gisaid_geoLocationRules.tsv
*/Virgin Islands/St. Thomas/ North America/USA/Virgin Islands/St. Thomas
North America/USA/Vi/* North America/USA/Virgin Islands/*
North America/USA/Virgin Islands of the U.S./* North America/USA/Virgin Islands/*
North America/USA/U.S. Virgin Islands/* North America/USA/Virgin Islands/*
North America/USA/Us Virgin Islands/* North America/USA/Virgin Islands/*
North America/USA/Virgin Islands Of The U.S./* North America/USA/Virgin Islands/*
North America/Caribbean/British Virgin Islands/ Europe/United Kingdom/British Virgin Islands/
North America/British Virgin Islands/*/* Europe/United Kingdom/British Virgin Islands/*
North America/U.S. Virgin Islands// North America/USA/Virgin Islands/
North America/U.S. Virgin Islands/U.S. Virgin Islands/ North America/USA/Virgin Islands/
North America/U.S. Virgin Islands/St. Thomas/ North America/USA/Virgin Islands/St. Thomas
North America/U.S. Virgin Islands/St. Croix/ North America/USA/Virgin Islands/St. Croix
South America/British Virgin Islands// Europe/United Kingdom/British Virgin Islands/
South America/British Virgin Islands/British Virgin Islands/ Europe/United Kingdom/British Virgin Islands/
South America/British Virgin Islands/Caribbean/ Europe/United Kingdom/British Virgin Islands/
If we want to special case zika to use USVI, I think this can be done with the local geolocation rule:
North America/USA/Virgin Islands/* North America/USVI/USVI/*
There was a problem hiding this comment.
Is my understanding correct that we want to label all "Virgin Islands" records as "USVI" not just the ones associated with this https://github.com/blab/zika-usvi/ study?
If so, is there a reason the zika build should use "USVI" while ncov, mpox, and other pathogens use "Virgin Islands"? I'm happy to add it here, just thinking through if this is a good design decision.
Alternatively, I'm happy to reverse this change and use "Virgin Islands" instead, as suggested by the 2nd solution here as being the better solution going forward.
- Change the spiked-in sequences to "Virgin Islands" to match GenBank.
There was a problem hiding this comment.
Yes, that's my understanding based upon the language used in the paper and the country strings we were using while maintaining the Zika tree during the outbreak.
There was a problem hiding this comment.
Thanks all! Switched to the local geolocation rule, and tried to summarize the rational in the commit message: 927222f
Please let me know if you catch anything else
j23414
force-pushed
the
40-use-a-consistent-country-name-across-usvi-samples-both-submitted-to-genbank-and-not-yet
branch
from
6215812 to
927222f
Compare
Label all "Virgin Islands" records as "USVI" to be consistent with text from publication: Black, A. et al., 2017. Genetic characterization of the Zika virus epidemic in the US Virgin Islands. bioRxiv, p.113100. https://www.biorxiv.org/content/10.1101/113100v2
j23414
force-pushed
the
40-use-a-consistent-country-name-across-usvi-samples-both-submitted-to-genbank-and-not-yet
branch
from
927222f to
04387cd
Compare
j23414
deleted the
40-use-a-consistent-country-name-across-usvi-samples-both-submitted-to-genbank-and-not-yet
branch
Description of proposed changes
Use a consistent country name across usvi samples both submitted to genbank and not yet.
Related issue(s)
Checklist