Example Bioinformatics application in Neo4j

Toni Hermoso Pulido

isnobig, BCN - 2015

@toniher

Objectives

• Store
• Relate
• Retrieve
• Calculate
• ... take decisions!

Directed Acyclic Graphs (DAGs)

NCBI Taxonomy - Gene Ontology (3 DAGs)

Simple Graph concepts

• Node
• Relationship

Neo4j

• Easy Installation
  • Java based
• REST API

Neo4j organisation concepts

• 1 server instance -> 1 database
• Labels can allow having multiple non-linked graphs
  • In our case we used: GO_TERM, TAXID
• Handling multiple instances: Neobox

Data import with py2neo

• py2neo: Python wrapper
• Based on Neo4j REST API
• Offers Batch import
• Recommendation:
  • Using pandas library (CSV)

Data import with py2neo

• Base Files: NCBI Taxonomy
• Base Files: Gene Ontology
• Code

Querying the graphs

Cypher

Graph Gist examples

Problems:

Shortest path between 2 nodes and Lowest Common Ancestor

Explanation

Solving problem with Cypher

Py2neo code example for NCBI Taxonomy

Solving problem with JAVA libraries

Example code

Putting everything in an unmanaged extension

• Server Configuration in conf
• Plugin creation
  • Maven (Java Package Manager)
    • pom.xml (definition and dependencies)
  • To be placed in plugins directory
• Extra libraries
  • e. g. minimal-json (placed in system/lib)

Extension as a REST service

• REST API Landing path (conf/neo4j-server.properties)

org.neo4j.server.thirdparty_jaxrs_classes=cat.cau.neo4j.biorelation.rest=/biodb

• Using JERSEY for REST API: javax.ws.rs
  • Similar to NodeJS Express, Python Flask or Perl Mojolicious
  • Documentation

Extension as a REST service

@Path("/parent")

        @GET
        @Path("/helloworld")

        @GET
        @Path("/distance/go/{acc1}/{acc2}")

Code

Questions?