Cambridge Healthtech Institute’s Inaugural

AI for Drug Discovery and Development

Accelerating Drug Discovery- One Use Case At a Time

June 2-4, 2020


Artificial Intelligence (AI), especially deep learning and machine learning, is coming out as disruptive technology for the faster discovery and development of innovative therapies. There is a lot of excitement about the opportunities associated with the application of AI, but at the same time, a gap exists in understanding these possibilities and applying them to drug discovery and development processes. CHI’s inaugural AI for Drug Discovery and Development conference will address the key questions such as: What can AI and ML do and not do for the pharmaceutical industry? What should be done to harness value out of AI for drug discovery? What measures should be taken to invest and apply AI at various stages of drug development, such as drug design, optimization safety prediction, CMC, quality control, clinical trials, repurposing, and business strategies? What should be the expectation of returns?

Final Agenda

Recommended Short Course*

SC6: An ML/AI Tutorial: From Basics to Advanced - Detailed Agenda

*Separate registration required.

Tuesday, June 2

10:00 am Main Conference Registration Open

CHALLENGES AND OPPORTUNITIES

11:15 Chairperson’s Remarks

Amol Jadhav, PhD, Industry Consultant, Transformational Health, Frost & Sullivan

11:25 KEYNOTE PRESENTATION: Using AI Tools to Accelerate Drug Discovery

hop_corneliusCornelis Hop, Vice President, Drug Metabolism & Pharmacokinetics, Genentech

This presentation will delve into the use of Machine Learning- and Artificial Intelligence-based applications in discovery and development projects. A sampling of what will be discussed: a retrospective analysis on predicting potency in drug discovery, use case data from current and past ADMET projects, and external collaborations to establish the benefits of these approaches.

11:55 Human Genetics-Based Drug Discovery: Challenges and Opportunities

Gavva_NarenderNarender Gavva, PhD, Director, Early Target Discovery, Takeda California, Inc.

The drug discovery industry adapted patient genetics target identification and validation (TIDVAL) approaches last decade to increase success rates in the clinic. There remain many challenges for human genetics TIDVAL in finding large effect size targets that can prevent or reverse disease progression. The presentation will cover opportunities for longitudinal studies that couple AI for drug discovery.

12:25 pm Artificial Intelligence Approach to Ligand and Structure-Based Design

enyedy_istvanIstvan Enyedy, PhD, Principal Scientist, Medicinal Chemistry, Biogen

Ligand and structure-based methods in combination with machine learning models are necessary components of a drug discovery campaign. We can increase the efficiency of optimizing compounds by combining these methods into a multiparameter optimization platform that combines all three approaches. Preliminary results of this approach will be presented.

12:55 Transition to Lunch

1:00 Luncheon Presentation (Sponsorship Opportunity Available) or Enjoy Lunch on Your Own

1:30 Session Break

USE CASES

2:00 Chairperson’s Remarks

Narender Gavva, PhD, Director, Early Target Discovery, Takeda California, Inc.

2:05 AI and the Cloud: Novel Ways to Accelerate InnovatioN

Todd Neuville, Leader, Worldwide Business Development, LeaderLife Sciences, Amazon Web Services (AWS)

Learn how pharma companies are working with artificial intelligence and machine learning (AI/ML) to accelerate research, enhance their clinical trials, improve manufacturing, and better understand real-world data. Hear how cloud technology is helping to expand the use of AI along the life sciences value chain to accelerate time to market for new products and increase operational efficiency.

2:35 A Deep Learning Approach to Antibiotic Discovery

Stokes_JonathanJonathan Stokes, PhD, Banting Fellow, Collins Lab, Broad Institute of MIT & Harvard

To address the antibiotic-resistance crisis, we trained a deep neural network to predict new antibiotics. We performed predictions on multiple chemical libraries and discovered a molecule from the Drug Repurposing Hub – halicin – that is structurally divergent from conventional antibiotics and displays activity against a wide spectrum of pathogens. Halicin also effectively treated Clostridioides difficile and Acinetobacter baumannii infections in mice. Deep learning approaches have utility in expanding our antibiotic arsenal.

3:05 Fringing – Or, How to Best Search for Gold Nuggets

Springer_ClaytonClayton Springer, PhD, Computational Chemist, Global Discovery Chemistry, Novartis Institutes for BioMedical Research, Inc.

The Fringing approach is inspired by Kriging. Kriging is a method from geostatistics which estimates the most likely distribution of gold based on samples from a few boreholes. Fringing translates this approach to chemical space and allows algorithmic exploitation and exploration of the chemical space.

3:35 Sponsored Presentation (Opportunity Available)

4:05 Networking Refreshment Break and Transition to Keynote


PLENARY KEYNOTE SESSION

4:25 - 6:05 Driving Entrepreneurial Innovation to Accelerate Therapeutic Discoveries

The life sciences community has an unprecedented scientific arsenal to discovery, develop and implement treatments, cures and preventions that enhance human healthcare.

Moderator: Nadeem Sarwar, President, Eisai Center for Genetics Guided Dementia Discovery (G2D2), Eisai Inc.

Panelists: Anthony Philippakis, Chief Data Officer, Broad Institute; Venture Partner, GV

Barbara Sosnowski, Vice President and Global Head, Emerging Science & Innovation Leads, WWRDM, Pfizer

John Hallinan, Chief Business Officer, Massachusetts Biotechnology Council

6:05 Welcome Reception in the Exhibit Hall with Poster Viewing

7:10 Close of Day

Wednesday, June 3

7:30 am Registration Open and Morning Coffee

AI STRATEGIES IN CLINICAL TRIALS

8:10 Chairperson’s Remarks

Janaki Iyer, Team Lead/Senior Medical Writer, INVIVO Communications, Inc..

8:15 KEYNOTE PRESENTATION: AI for Acceleration of Drug Development

John_BinoBino John, PhD, Associate Director, Data Science, Clinical Pharmacology and Safety Sciences – Data Science and AI, AstraZeneca

Drug development is an expensive and costly endeavor, costing an average of 2.6 billion dollars to bring a drug to market. Artificial Intelligence is essential in reducing the costs and time to bring these to the clinic. This talk will highlight some of the current AI initiatives at AstraZeneca, spanning chemical and biological data use cases that seek to improve drug design and develop safer medicines.

8:45 AI and ML Approaches to Healthcare Data Integration and Analysis

Bharadwaj_ShruthiShruthi Bharadwaj, PhD, Senior Scientist, Novartis Oncology Precision Medicine

With the increase in availability of clinical trial data, AI and Machine Learning Approaches are becoming imperative in mining and finding clinically significant insights. In this talk, I will provide an overview of the various approaches currently used to tackle the big-data problem in pharma.

9:15 Boosting Clinical Trial Success Rates with AI Strategies

Iyer_JanakiJanaki Iyer, Team Lead/Senior Medical Writer, INVIVO Communications, Inc.

The typical drug discovery and development process, commonly termed as “bench to bedside”, lasts about 10-15 years and costs over a billion dollars. Failed clinical trials can lead to tremendous losses in terms of both time and money. This talk will discuss AI strategies as a viable option to enhance trial designs, improve patient recruitment strategies, and advance patient monitoring with the aim of maximizing overall clinical trial success rates.

9:45 Sponsored Presentation (Opportunity Available)

10:15 Coffee Break in the Exhibit Hall with Poster Viewing

11:00 Strategies for Building AI-Ready Data Sources and (Semi)Autonomous Reasoning Agents Operating on Top of Them

Grotthuss_Marcin_vonMarcin von Grotthuss, PhD, Senior Computational Scientist, Broad Institute of Massachusetts Institute of Technology and Harvard

Here, we present a prototype Translator framework and architecture, which we have developed for integrating semantically, annotated Knowledge Sources (over 40) and for creating a data platform to support automated reasoning and serendipitous discovery of new ‘facts’ or interesting and testable hypotheses. We also discuss the strategies of how to integrate and provide high-value AI-ready data sources as well as how to develop (semi) autonomous reasoning agents that would advance reasoning through innovative uses of these knowledge sources.

AI IN PREFORMULATION STUDIES

11:30 Segmentation and Classification of Crystalline Structures from 3D X-Ray Microscopy Images in Pharmaceutical Tablets

Babburi__PradeepPradeep Babburi, MS, Data Scientist, R&D, AbbVie, Inc.

Here we present ongoing work using image analysis and machine/deep learning techniques to segment and differentiate the crystalline and amorphous phases of the drug as well as other crystalline substances like silicon dioxide (SiO2) from 3D x-ray microscopy (XRM) scans. Our work demonstrates the results of basic image analyses, geometric feature extraction, as well as unsupervised and supervised learning models trained to identify the crystalline structures based on their morphology.

12:00 pm Sponsored Presentation (Opportunity Available)

12:30 Transition to Lunch

12:35 Luncheon Presentation (Sponsorship Opportunity Available) or Enjoy Lunch on Your Own

1:05 Session Break


PLENARY KEYNOTE SESSION

1:45 - 3:15

Lgr5 Stem Cell-Based Organoids in Human Disease

Clevers_HansHans Clevers, MD, PhD, Principal Investigator of Hubrecht Institute and Princess Máxima Center, CSO of HUB Organoids Technology

Organoid technology opens a range of applications in fields such as physiology, study of disease, drug development and personalized medicine. Human organoids represent excellent disease models, be it infectious, hereditary or malignant  Eventually, cultured mini-organs may be used to replace transplant organs from donors. I will describe how we originally created ‘mini-guts’ via 3D culture systems of stem cells of the small intestine and colon, and then expanded the technology to virtually all human organs.

Systematically Drugging Ras

Fesik_StephenStephen Fesik, PhD, Professor of Biochemistry, Pharmacology, and Chemistry, Orrin H. Ingram II Chair in Cancer Research, Vanderbilt University School of Medicine

K-Ras is a small GTPase that is mutated in pancreatic (90%), colon (50%), and lung (30%) carcinomas. Downregulation of activated Ras reverses the transformed phenotype of cells and results in the dramatic regression of tumors in murine xenograft models. Thus, K-Ras inhibition represents an attractive therapeutic strategy for many cancers. In this presentation, I will discuss our efforts to directly target Ras at two sites and target SOS, a molecular partner of Ras, with activators and inhibitors. 

3:15 Refreshment Break in the Exhibit Hall with Poster Viewing

USE OF AI/ML FOR ADME/Tox PREDICTIONS

4:00 Chairperson’s Remarks

Barun Bhhatarai, PhD, Investigator, Novartis Institute for Biomedical Research

4:05 ML and AI on ADME/Tox Accelerating Drug Discovery

Bhhatarai_BarunBarun Bhhatarai, PhD, Investigator, Novartis Institute for Biomedical Research

ML- and AI-related approaches have been tested and applied in various areas within Novartis. In ADMETox, ML approaches are serving intended purposes and complementing experimental methods. With the advent of AI, ingenious deep learning algorithms, and powerful micro-processors, we have explored its anticipated benefit in preclinical and clinical programs. Our various efforts on data digitization, ML and AI implementation, and collaborations will be discussed with specific examples from ADMETox.

4:35 Artificial Intelligence and Small-Molecule Drug Metabolism

Swamidass_JoshuaJoshua Swamidass, MD, PhD, Assistant Professor, Immunology and Pathology, Laboratory and Genomic Medicine; Faculty Lead, Translational Informatics, Institute for Informatics, Washington University

We have been building artificial intelligence (AI) models of metabolism and reactivity. Metabolism can both render toxic molecules safe and safe molecules toxic. The AI models we use quantitatively summarize the knowledge from thousands of published studies. The hope is that we could more accurately model the properties of medicines to determine whether metabolism renders drugs toxic or safe. This is one of many places where artificial intelligence could give traction on the difficult questions facing the industry.

5:05 Find Your Table, Meet Your Moderator

5:10 Roundtable Breakout Discussions - View Details

TABLE: Decoding AI: Making the Case for Artificial Intelligence in the Pharma Industry

Moderator: Amol Jadhav, PhD, Industry Consultant, Transformational Health, Frost & Sullivan

TABLE: Machine Learning in Action: Moving Beyond Hype

Moderator: Sean Ekins, PhD, DSc, CEO, Collaborations Pharmaceuticals, Inc.

5:45 Reception in the Exhibit Hall with Poster Viewing

6:45 Close of Day

Thursday, June 4

8:00 am Registration Open and Morning Coffee


PLENARY KEYNOTE SESSION

8:30 - 9:40 Applications of Artificial Intelligence in Drug Discovery – Separating Hype from Utility

Walters_PatrickPatrick Walters, PhD, Senior Vice President, Computation, Relay Therapeutics

Over the last few years, there has been tremendous interest in the application of artificial intelligence and machine learning in drug discovery. Ultimately, the success of any predictive model comes down to three factors: data, representation, and algorithms. This presentation will provide an overview of these factors and how they are critical to the successful implementation and deployment of AI methods.

9:40 Coffee Break in the Exhibit Hall with Poster Viewing

10:25 Chairperson’s Remarks

Sean Ekins, PhD, DSc, CEO, Collaborations Pharmaceuticals, Inc.

10:30 PANEL DISCUSSION: Challenges in Adoption and Implementation: Hype, Trust, Privacy, and Explainable AI

Ekins_SeanModerator: Sean Ekins, PhD, DSc, CEO, Collaborations Pharmaceuticals, Inc.


Panelists:

Joseph Lehar, PhD, Vice President, Data Science, The Janssen Pharmaceuticals

Walters_PatrickPatrick Walters, PhD, Senior Vice President, Computation, Relay Therapeutics


Jadhav_AmolAmol Jadhav, PhD, Industry Consultant, Transformational Health, Frost & Sullivan


Jonathan Lefman, PhD, Developer Relations Manager, Healthcare and Life Sciences, Nvidia

John_BinoBino John, PhD, Associate Director, Data Science, Clinical Pharmacology and Safety Sciences – Data Science and AI, AstraZeneca


Gavva_NarenderNarender Gavva, PhD, Director, Early Target Discovery, Takeda California, Inc.


AI IN R&D STRATEGY AND BUSINESS DECISION

11:30 Achieving Digital Disruption in Pharma through Artificial Intelligence – Status & Opportunities

Jadhav_AmolAmol Jadhav, PhD, Industry Consultant, Transformational Health, Frost & Sullivan

This presentation will focus around the commercial aspects and call out major current application areas of AI in the pharmaceutical industry. Opportunity assessment within specific sub-segments, themes driving adoption, preview of successful business models and relevant case studies, expectation on returns, and global scenarios will be discussed.

12:00 pm Sponsored Presentation (Opportunity Available)

12:30 Transition to Lunch

12:35 Luncheon Presentation (Sponsorship Opportunity Available) or Enjoy Lunch on Your Own

1:05 Dessert and Coffee Break in the Exhibit Hall with Poster Viewing

2:00 Chairperson’s Remarks

Jonathan Lefman, PhD, Developer Relations Manager, Healthcare and Life Sciences, Nvidia

2:05 Application of AI in Pharma R&D: Use Cases

Jonathan Lefman, PhD, Developer Relations Manager, Healthcare and Life Sciences, Nvidia

2:35 Building a Small Company to Apply Machine Learning for Rare and Neglected Disease Drug Discovery

Ekins_SeanSean Ekins, PhD, DSc, CEO, Collaborations Pharmaceuticals, Inc.

Collaborations Pharmaceuticals, Inc. (CPI) aims to streamline the development of drugs for rare and neglected tropical diseases. Using our machine learning technology and combining forces with many academic collaborators, we have identified treatments for parasites (T. cruzi), bacteria (M. tuberculosis), and viruses (Ebola, HIV, etc.), progressing to in vivo models. I will describe how we can also apply this approach for rare diseases.

3:05 Application of DL Approaches for Non-Target-Based Drug Repurposing

Arshadi_Arash_KeshavarziArash Keshavarzi Arshadi, MS, Research Fellow, College of Medicine, University of Central Florida

We talk about the use of DL approaches – especially transfer learning – for predicting the potency of already approved drugs for other diseases. Since the target of the known drugs would be completely different types of biomolecules in different cells, pursuing drug repositioning with target-based approaches would not be applicable. Also, many target molecules or mechanisms of their interactions are not discovered yet. Therefore, non-target approaches would be suitable for this manner. In the case of having low data, we will discuss how transfer learning would increase accuracy and recall.

3:35 Close of Conference