Postgresql error for isolate_name dose not match what actually happened

[KaimingTao]

Paper: Boschi22
Isolate name: hCoV-19/France/PAC-IHUMI-2096/2020
Variant name: B.1.160

The paper is about mAbs, not plasmas, but the error prompt is about the subject history of plasmas.

Screenshot

Currently, these records are excluded from the database.

[philiptzou]

This is not a bug. It's a validation I added to forbid using B.1.160 (477N+614G) as the var_name of an isolate which is linked to at least one susceptibility test result. Leave the var_name "NULL" can solve this problem.

[KaimingTao]

This is not a bug. It's a validation I added to forbid using B.1.160 (477N+614G) as the var_name of an isolate which is linked to at least one susceptibility test result. Leave the var_name "NULL" can solve this problem.

Sorry I didn't get the context. Why B.1.160 can't be used for an isolate, and the isolate is used in the susc_result table?

Also, the issue is not about the var_name. The prompt mentioned subject_history table, but the B.1.160 used in Boschi22 is only about monoclonal antibodies.

[philiptzou]

@KaimingTao It's the same reason why we don't assign D.2 to an isolate. You can find the detailed reason by searching for and read the email thread "Why is S477N mutation listed on the dashboard as D2 (S477N)". You can only use B.1.160 and D.2 in subject_history.

Basically, D.2 and B.1.160 only includes two Spike mutations S477N+D614G. We decided to aggregate the neutralization result of D.2 and B.1.160 by their mutation pattern but not by their variant so the results can be merged with S477N. However, if you specify a variant to the isolate, the neutralization result will be aggregated by its variant but not by its mutation pattern.

You can still specify B.1.160 or D.2 to an isolate used as infected virus. Since 1) #11 will eventually change infected_isolate to infected_variant in the proposed subject_infections table; 2) infected viruses don't use the same complicated aggregation rule used by control and test isolates - they always aggregated by their variants.

[philiptzou]

This issue is moved to covid-drdb since it's related to this constraint:

covid-drdb/constraints_post-import.sql

Lines 4 to 23 in 4732b8d

	DO $$
	DECLARE row isolates%rowtype;
	BEGIN
	FOR row IN SELECT * FROM isolates iso WHERE
	var_name IN ('B.1.177.75', 'B.1.160', 'D.2') AND (
	EXISTS (
	SELECT 1 FROM rx_potency pot WHERE
	pot.iso_name=iso.iso_name
	) OR
	EXISTS (
	SELECT 1 FROM rx_fold f WHERE
	f.control_iso_name=iso.iso_name OR
	f.iso_name=iso.iso_name
	)
	)
	LOOP
	RAISE EXCEPTION E'Isolate of variant \x1b[1m%\x1b[0m can be only used as an infected isolate in subject_history. Unlink \x1b[1m%\x1b[0m from \x1b[1m%\x1b[0m to be used in rx_potency or rx_fold.', row.var_name, row.iso_name, row.var_name;
	END LOOP;
	END
	$$;