Neo4j Demo: Using Knowledge Graphs to Classify Diabetes Patients (GlaxoSmithKline)
0 likes220 views
This document discusses using Neo4j's graph data science capabilities to classify diabetes patients based on transcriptomics data and connect patients to knowledge graphs. It demonstrates loading patient data containing transcripts measured for each patient, generating graph embeddings to represent the data while preserving topology, training a node classification model to predict diabetes status, and performing community detection for subphenotyping. The goal is to transform heterogeneous clinical data into an integrated graph representation to power machine learning and connect insights to existing domain knowledge.
Neo4j Demo: Using Knowledge Graphs to Classify Diabetes Patients (GlaxoSmithKline)
- 1. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. 1 Neo4j Demo Classify Diabetes Patients and Connect to Knowledge Graphs Dr Alexander Jarasch Field Engineering Specialist Pharma / Healthcare / Biotech
- 2. 2 Use Cases for the Entire Drug Life Cycle Target Discovery Hit Generation Lead Identification Lead Optimization Animal Models Clinical Trials FDA/EMA Review & Approval Post Approval Manufacturing F100 F100 Cyber security F100
- 3. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. 3 Native Graph Database The foundation of the Neo4j platform; delivers enterprise-scale and performance, security, and data integrity for transaction and analytical workloads. Data Science and Analytics Explorative tools, rich algorithm library, and Integrated supervised Machine Learning framework. Development Tools & Frameworks Tooling, APIs, query builder, multi-language support for development, admin, modelling, and rapid prototyping needs. Discovery & Visualization Code-free querying, data modeling and exploration tools for data scientists, developers, and analysts. Graph Query Language Support Cypher & openCypher; Ongoing leadership and standards work (GQL) to establish lingua franca for graphs. Ecosystem & Integrations Rich ecosystem of tech and integration partners. Ingestion tools (JDBC, Kafka, Spark, BI Tools, etc.) for bulk and streaming needs. Runs Anywhere Deploy as-a-Service (AuraDB) or self-hosted within your cloud of choice (AWS, GCP, Azure) via their marketplace, or on-premises. Neo4j Graph Data Platform
- 4. © 2022 Neo4j, Inc. All rights reserved. 4 65+ Graph Algorithms - Out of the Box Pathfinding & Search Centrality Community Detection ❏ Delta-Stepping Single-Source ❏ Dijkstra’s Single-Source ❏ Dijkstra Source-Target ❏ All Pairs Shortest Path ❏ A* Shortest Path ❏ Yen’s K Shortest Path ❏ Minimum Weight Spanning Tree ❏ Random Walk ❏ Breadth & Depth First Search ❏ Degree Centrality ❏ Closeness Centrality ❏ Harmonic Centrality ❏ Betweenness Centrality & Approx. ❏ PageRank ❏ Personalized PageRank ❏ ArticleRank ❏ Eigenvector Centrality ❏ Hyperlink Induced Topic Search (HITS) ❏ Influence Maximization (Greedy, CELF) ❏ Weakly Connected Components ❏ Strongly Connected Components ❏ Label Propagation ❏ Leiden ❏ Louvain ❏ K-Means Clustering ❏ K-1 Coloring ❏ Modularity Optimization ❏ Speaker Listener Label Propagation ❏ Approximate Max K-Cut ❏ Triangle Count ❏ Local Clustering Coefficient ❏ Conductance Metric Heuristic LP Similarity Graph Embeddings ❏ Adamic Adar ❏ Common Neighbors ❏ Preferential Attachment ❏ Resource Allocations ❏ Same Community ❏ Total Neighbors ❏ K-Nearest Neighbors (KNN) ❏ Filtered K-Nearest Neighbors (KNN) ❏ Node Similarity ❏ Filtered Node Similarity ❏ Similarity Functions ❏ Fast Random Projection (FastRP) ❏ Node2Vec ❏ GraphSAGE ❏ Node Classification Pipeline ❏ Link Prediction Pipeline ❏ Node Regression Pipeline
- 5. © 2022 Neo4j, Inc. All rights reserved. 5 What We Do in This Demo • GDS Node Classification Pipeline on patient data • Loading data • Data preparation -> Fast Random Projections • Model training • Predict T2D / non-T2D based on transcriptomics • GDS Community Detection • For Subphenotyping
- 6. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. From Chaos to Structure: Neo4j Graph Data Science is Changing How Machine Learning Gets Done Graph Embeddings summarize the enhanced explicit knowledge of a graph 6
- 7. © 2022 Neo4j, Inc. All rights reserved. 7 Dataset
- 8. © 2022 Neo4j, Inc. All rights reserved. 8 Dataset • 63 patients (total) • 51 classified as T2D / non-T2D • 12 unclassified • 18k transcripts measured on each patient
- 9. © 2022 Neo4j, Inc. All rights reserved. 9 Data Model Transcript Patient :MEASURED value :SIMILAR similarityScore
- 10. © 2022 Neo4j, Inc. All rights reserved. 10 Demo Time
- 11. © 2022 Neo4j, Inc. All rights reserved. 11 Coming Back to Knowledge Graphs
- 12. © 2022 Neo4j, Inc. All rights reserved. 12 Clinical Data from Patients Gender Age BMI diab.status Sample coll #samples #used surgery disease Histological diagnosis
- 13. Neo4j, Inc. All rights reserved 2022 13 Transform Data with GDS - Fast Random Projections CALL gds.fastRP.write( 'patients', { embeddingDimension: 50, writeProperty: ‘fastRP- embedding’ } ) YIELD nodePropertiesWritten
- 14. Neo4j, Inc. All rights reserved 2022 14 Connect Patient Data with Your Knowledge Graph Performance Increase 20 - 12’000x
- 15. © 2022 Neo4j, Inc. All rights reserved. 15 Summary • Graph Data Science library to perform out of the box machine learning • Using Graph Embeddings (FastRP) to represent heterogenous data as vector keeping the topology • New GDS pipeline for Node Classification • GDS Community Detection
- 16. © 2022 Neo4j, Inc. All rights reserved. © 2022 Neo4j, Inc. All rights reserved. 16 Thank you! Contact me alexander.jarasch@neo4j.com