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
@nunogit , @mkutmon and me just discussed the webservice of bridgeDb in more detail.
We believe that if we add more documentation, and specific examples for gene, metabolite and interaction queries (per call), the webservice could be vidualised on a single list, without having the need to split them for Genes and metabolites (as is currently the case).
This split is a hack, because the two calls now listed for metabolites are exactly the same as the gene calls. However, it could look to users that there are only 2 calls possible now for metabolites (while actually all gene calls should work for metabolites too).
I'm not sure why the decision has been made to split the list of calls up between genes and metabolites, but if we want to include interactions (and at a later stage the gene-variants from Freddie, and in the future other mapping datasets), we need to think of a good solution for the future.
Nuno would like to check if it is possible to get clickable examples, which then populate the call.
See the artist impression below:
The text was updated successfully, but these errors were encountered:
egonw
transferred this issue from bridgedb/bridgedb-webservice
Feb 26, 2023
Denise wrote in https://github.com/BiGCAT-UM/BiGCaT-ICT/issues/138 (now closed, and continued here):
@nunogit , @mkutmon and me just discussed the webservice of bridgeDb in more detail.
We believe that if we add more documentation, and specific examples for gene, metabolite and interaction queries (per call), the webservice could be vidualised on a single list, without having the need to split them for Genes and metabolites (as is currently the case).
This split is a hack, because the two calls now listed for metabolites are exactly the same as the gene calls. However, it could look to users that there are only 2 calls possible now for metabolites (while actually all gene calls should work for metabolites too).
I'm not sure why the decision has been made to split the list of calls up between genes and metabolites, but if we want to include interactions (and at a later stage the gene-variants from Freddie, and in the future other mapping datasets), we need to think of a good solution for the future.
Nuno would like to check if it is possible to get clickable examples, which then populate the call.
See the artist impression below:
The text was updated successfully, but these errors were encountered: