← Projects

kegg-cli

A command-line pathfinder over the KEGG biochemical database — depth-first search for the shortest reaction path between two compounds.

  • Python

A cross-platform CLI for querying the Kyoto Encyclopedia of Genes and Genomes (KEGG) database: find, list, and get information on any compound, enzyme, or reaction, and — the part I worked on — discover a reaction pathway between two compounds that aren’t directly connected.

The pathfinding

Two depth-first search variants solve for a route between a starting and target compound: one searches over KEGG’s own curated modules (pre-built pathways), the other searches directly over individual reactions when no curated module connects the two compounds. Same classical graph-search idea used for shortest-path problems generally, applied here to a biochemical reaction network instead of a road network or a state graph.

Built at Cornell as a bioengineering course project (BEE 3600) with Shaumik Ashraf, under Prof. Buzz Barstow — the repo lives under my collaborator’s account since it was a joint project.