Add KG subset operation to KGX

biolink / kgx

KGX is a Python library for exchanging Knowledge Graphs

https://kgx.readthedocs.io

BSD 3-Clause "New" or "Revised" License

115 stars 27 forks source link

Add KG subset operation to KGX #237

Open deepakunni3 opened 3 years ago

deepakunni3 commented 3 years ago

Add a KG subset/extract operation similar to Robot extract.

cmungall commented 3 years ago

Use cases:

KG-COVID-19, include subset of NCBI taxonomy of relevance (e.g. coronaviruses and their hosts), no need to include all archaea
Include only small molecules of relevance from a drug/molecule database, plus anything connected to them

the algorithm would be different. Here we typically want to start with a seed and include anything reachable. Robot uses SLME or MIREOT. Although I think we want the reachability algorithm in ROBOT, cc @matentzn

matentzn commented 3 years ago

I think reachability is the most natural here for the KG case; should be trivial with networkx or whatever graph representation you have internally!

For ROBOT, yeah, that's a bit harder; thanks for placing it in my mind again.