For context, one of the most significant changes between OTG and the new Platform is the definition of a variant in the variant index. Whereas we used to include every variant in GnomAD (~70M), we now only include variants for which we have any phenotypic information available (~6.5M). Some resources (e.g. Ensembl or GnomAD) now contain variants over 700M, so it's expected that a user might land on the Platform with a variant we don't have. For this reason, it would be important that the user understands why before they give up or start bombarding us with issues.
Because variants can have predictable identifiers, it could be beneficial to anticipate when a user is trying to request a variant and we have no results to trigger a message explaining why no variant was found.
The patterns for variants could be of two different types:
RSID The user is trying to find an rsid. The pattern would start with rs followed by a variable number of integers (e.g. rs4129267)
Chr_pos_ref_alt This is our primary identifier but can have a few variations. I will break it down:
chr: it can be anything in the range 1-22, X, Y or MT
_: this separator can be _ or :
pos: a positive integer
_: this separator can be _ or :
ref: this is a string of variable size in the range [A-Za-z]
_: this separator can be _ or -
alt: this is a string of variable size in the range [A-Za-z]
Regarding the message. I'd like to suggest the following text as a placeholder. We can adjust the text later on:
**Variant not found**. The Open Targets Platform only includes variants with a known phenotypic association. For more information, see [documentation](https://platform-docs.opentargets.org/).
I'm assigning it to you, @carcruz, in the first place, and we can all agree on how and when to do this.
Please note: for PPP 24.12 release, the warning message should point to PPP documentation, where we should make sure we explain our 'variants' inclusion choices
For context, one of the most significant changes between OTG and the new Platform is the definition of a variant in the variant index. Whereas we used to include every variant in GnomAD (~70M), we now only include variants for which we have any phenotypic information available (~6.5M). Some resources (e.g. Ensembl or GnomAD) now contain variants over 700M, so it's expected that a user might land on the Platform with a variant we don't have. For this reason, it would be important that the user understands why before they give up or start bombarding us with issues.
Because variants can have predictable identifiers, it could be beneficial to anticipate when a user is trying to request a variant and we have no results to trigger a message explaining why no variant was found.
The patterns for variants could be of two different types:
RSID The user is trying to find an rsid. The pattern would start with
rs
followed by a variable number of integers (e.g.rs4129267
)Chr_pos_ref_alt This is our primary identifier but can have a few variations. I will break it down:
chr
: it can be anything in the range 1-22, X, Y or MT_
: this separator can be_
or:
pos
: a positive integer_
: this separator can be_
or:
ref
: this is a string of variable size in the range [A-Za-z]_
: this separator can be_
or-
alt
: this is a string of variable size in the range [A-Za-z]Regarding the message. I'd like to suggest the following text as a placeholder. We can adjust the text later on:
I'm assigning it to you, @carcruz, in the first place, and we can all agree on how and when to do this.