Zichen Wang

[introductory/intermediate] Graph Machine Learning for Healthcare and Life Sciences

Summary

Graphs (or networks) are ubiquitous representations in life sciences and medicine, from molecular interactions maps, signaling transduction pathways, to graphs of scientific knowledge and patient-disease-intervention relationships derived from population studies and electronic health records. Recent advances in graph machine learning (ML) approaches such as graph neural networks (GNNs) have transformed a diverse set of problems relying on biomedical networks that traditionally depend on descriptive topological data analyses. In this lecture, I will first give an introduction to graph theory, ML algorithms and applications on graphs, particularly in the healthcare and life sciences domain. Then, we will dive deep into two areas of graph ML that have made significant contributions to the healthcare domain: 1) graph-level prediction tasks for molecules, and 2) knowledge graph generation and completion.

Syllabus

• Introduction to graph theory, network analysis, and graph ML
• Overview of graph ML in healthcare and life sciences
• Graph neural networks for small- and macro-molecules
• Organizing and generating new knowledge for healthcare with knowledge graphs

References

[1]       Li MM, Huang K, Zitnik M (2021) Graph Representation Learning in Biomedicine. https://arxiv.org/abs/2104.04883

[2]       Gaudelet T et al. (2021) Utilizing graph machine learning within drug discovery and development. Briefings in Bioinformatics, DOI: https://doi.org/10.1093/bib/bbab159

[3]       Li M et al. (2021) DGL-LifeSci: An Open-Source Toolkit for Deep Learning on Graphs in Life Science. ACS Omega. https://pubs.acs.org/doi/10.1021/acsomega.1c04017

[4]       Wang Z et al. (2021) LM-GVP: A Generalizable Deep Learning Framework for Protein Property Prediction from Sequence and Structure. bioRxiv https://www.biorxiv.org/content/10.1101/2021.09.21.460852v1

[5]       Wise C et al. (2020) COVID-19 knowledge graph: accelerating information retrieval and discovery for scientific literature. https://arxiv.org/abs/2007.12731

[6]       Zheng D et al. (2020) DGL-KE: Training knowledge graph embeddings at scale. SIGIR. https://arxiv.org/abs/2004.08532

Pre-requisites

Addressed to professionals, researchers and students who want to understand and apply graph ML in healthcare and life sciences. A basic understanding or interest in healthcare/biology is beneficial but not required.

Short bio

Zichen Wang is an applied scientist and the graph ML lead at Amazon ML Solutions Lab where he developed various graph ML applications for AWS customers in healthcare, life sciences, and other industries. He received his Ph.D. degree in Computational Biology from Icahn School of Medicine at Mount Sinai in New York, NY, USA in 2016. He continued his research in biomedical networks, systems pharmacology, and ML for healthcare as a postdoctoral fellow and a research-track assistant professor at Mount Sinai. He has made contributions in many areas of biomedical informatics including bioinformatics software development, drug discovery, functional genomics, clinical informatics, and protein function prediction. He published over 50 peer-reviewed articles (h-index of 32 with more than 12,000 citations) in prestigious journals and conferences such as Nature Medicine, Nature Communications, Nucleic Acids Research, and ICLR.