You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current metakb-cvc model doesn't use the Core IM representation of confidence/evidence level and direction properties in a way that I feel would be best to provide clear and consistent representations of statement semantics.
Re: direction - the ClinVar data example uses a bespoke/generic 'direction' attribute that is not found in the Core IM (which provides separate fields for 'confidence_direction' and 'evidence_direction').
Re: direction: the ClinVar data example does not bother using the 'confidence_level' attribute to report whether the statement reflects 'likely' or 'definitive' assessment of truth. The only place that these confidence assessments would appear is in the 'classification' field, which would contain values such as 'likely pathogenic'. This 'classification' field was sold to me as being complementary to the broken out specification of direction and level using core IM 'confidence_level' and 'confidence_direction' properties created for these semantics. I don't like the idea that the explicit/standard representation the VA team had previously agreed to be best (see #71) can be dropped in favor of this 'shorthand' property added to provide the user with a familiar community-specific term.
All of this speaks to the need to align thinking about / representation of evidence and confidence direction and level assessments.
The text was updated successfully, but these errors were encountered:
The current metakb-cvc model doesn't use the Core IM representation of confidence/evidence level and direction properties in a way that I feel would be best to provide clear and consistent representations of statement semantics.
Re: direction - the ClinVar data example uses a bespoke/generic 'direction' attribute that is not found in the Core IM (which provides separate fields for 'confidence_direction' and 'evidence_direction').
Re: direction: the ClinVar data example does not bother using the 'confidence_level' attribute to report whether the statement reflects 'likely' or 'definitive' assessment of truth. The only place that these confidence assessments would appear is in the 'classification' field, which would contain values such as 'likely pathogenic'. This 'classification' field was sold to me as being complementary to the broken out specification of direction and level using core IM 'confidence_level' and 'confidence_direction' properties created for these semantics. I don't like the idea that the explicit/standard representation the VA team had previously agreed to be best (see #71) can be dropped in favor of this 'shorthand' property added to provide the user with a familiar community-specific term.
All of this speaks to the need to align thinking about / representation of evidence and confidence direction and level assessments.
The text was updated successfully, but these errors were encountered: