eMetaboHUB / Forum-DiseasesChem

A Knowledge Graph from public databases and scientific literature to extract associations between chemicals and diseases.
Other
16 stars 6 forks source link

Integration of a Fuzzy-search algorithm for requesting FORUM by chemical/concept names #5

Closed maximeDelmas closed 2 years ago

maximeDelmas commented 3 years ago

Context


The current FORUM Web interface only allows to access associations using PubChem, ChEBI, ChemOnt or MeSH identifiers, with which users can be unfamiliar.

Desired behaviour


As a user, I could request for FORUM associations using chemical names from PubChem/ChEBI/ChemOnt and descriptor names from MeSH.

Solutions ?


As chemical names can be complex, a fuzzy-search algorithm seems to be a good choice for helping users find compounds of interest. As a first step only principal names should be included as the additional burden of the synonyms provided by PubChem (several millions) could lead to computational issues.

A first test on a fuzzy-search algorithm has been implemented here

@maximeDelmas @npaulhe

npaulhe commented 3 years ago

first proof of concept for fuzzy matching:

poc_fuzzy.zip

npaulhe commented 2 years ago

implemented in release 1.0.0https://forum-webapp.semantic-metabolomics.fr/