Deep Learning in PKPD and Pharmacometrics: Diversity of Approaches and Applications

Recorded On: 11/13/2024

Session Description

Description: Aritificial intelligence and machine learning are of increasing interest in the pharmacometrics space. Early efforts have focused on statistical machine learning approaches with emerging efforts in deep learning as well. Deep Learning encompasses a variety of approaches that present a breadth of opportunities. This session features three presentations describing the use of different deep learning approaches with diverse applications using qualitatively different data. Specifically, it will feature applications and associated technical detailsof protein language models, neural ODEs, and graph neural networks, Furthermore, the applications range from early antibody design, to preclinical PD/tumor modeling, to PK prediciton and generalization. The goal (learning objective) is to spur greater awareness, discussion, exploration, and use of a diversity of novel deep learning approaches across broad areas of pharmacometrics. All 3 speakers have been investing heavily in these approaches at different organizations and bring unique perspectives to the topic.

Conference Date

November 10-13, 2024

Conference Location

Phoenix, Arizona, USA

DOI

10.70534/LDPE3146

Omid Bazgir, PhD

Senior AI Scientist, Clinical Pharmacology M&S

Genentech Inc.

Omid Bazgir is a Senior AI Scientist in the Clinical Pharmacology department at Genentech. He received his Ph.D. in Electrical and Computer Engineering from Texas Tech University, focusing on developing machine learning algorithms for drug discovery and computational biology applications. At Bayer, Omid contributed to the advancement of computer vision tools, applying AI in the crop sciences industry. He joined Genentech in the summer of 2021 to develop machine learning and generative AI tools for drug development and clinical pharmacology applications. He has co-authored more than 20 journal and conference publications, white papers, and AI/ML patents.

Louis Joslyn, PhD

Principal Scientist

Preclinical and Translational Pharmacokinetics/Pharmacodynamics, Genentech

Louis Joslyn is a Principal Scientist in the Preclinical and Translational Pharmacokinetics/Pharmacodynamics department at Genentech. He specializes in developing mechanistic Quantitative Systems Pharmacology (QSP) models for various cancer immunotherapies, including T-cell therapies, T-cell engagers, antibody-drug conjugates (ADCs), and engineered cytokines. Additionally, Louis has expertise in creating machine learning frameworks for predicting pharmacokinetics (PK). He earned his Ph.D. in Computational Medicine and Bioinformatics from the University of Michigan, where he utilized systems modeling and virtual patient methodologies to characterize T-cell responses during tuberculosis and vaccination scenarios. Louis also holds a Bachelor of Arts in Mathematics and Computer Science from Simpson College. Broadly, Louis enjoys employing interdisciplinary approaches to explore and explain biological phenomena.

Sean Hayes, PhD

Associate Director, Data Science

Merck Research Laboratories

Sean Hayes is currently an Associate Director supporting AI/ML in Pharmacometrics at Merck Research Labs. Sean has been developing and supporting AI/ML methods for drug discovery and development at Merck for the last 6 years, starting from a postdoctoral research project. Sean has also provided modelling support for development programs accross therepeutic areas, particularly immunology, vaccines, and neuroscience. Sean originally trained as a theoretical ecologist, which initiated his broad interest in systems biology and novel computational approaches such as AI/ML. Currently Sean is focused on on identifying and implementing emerging AI/ML methods to both dramatically accelerate critical modelling support tasks and to enable novel pharmacometric capabilities.