2024 ARCHIVES
Sunday, May 12
Main Conference Registration1:00 pm
Recommended Pre-Conference Short Course2:00 pm
SC3: In silico and Machine Learning Tools for Antibody Design and Developability Predictions
*Separate registration required. See short course page for details.
Monday, May 13
Registration and Morning Coffee7:00 am
Chairperson’s Remarks
Timothy Patrick Jenkins, PhD, Assistant Professor & Head, Data Science, DTU Bioengineering
Enhancing Molecule Developability Studies with Automated Analytical Systems
Benjamin Weiche, PhD, Senior Scientist, Large Molecule Research, Biochemical & Analytical Resaerch, Roche Innovation Center Munich
Developability or molecule assessment represents a key capability across biopharmaceutical research and development to guide selection of the best clinical lead molecules. We present the use of cutting-edge automated analytical systems transforming efficiency and precision of developability studies and impacting drug development timelines.
High-Throughput Biophysical Characterization and Comprehensive Analysis
Ivan Budyak, PhD, Director, Analytical Development, Biophysical Characterization, Eli Lilly and Co.
High-throughput biophysical characterization and comprehensive analysis are essential to understanding colloidal and conformational behaviors of bioproducts. This presentation will discuss the use of high-throughput biophysical characterization techniques, implementation of data curation tools, and data analysis using machine learning algorithm to provide platform level understanding of colloidal, thermal, and chemical stability of bioproducts.
KEYNOTE PRESENTATION: Enabling Pipeline Programs Lead Candidate Selection & Optimization through Data Structure, Statistical Analysis & ML/AI
Marc Bailly, PhD, Principal Scientist, Protein Sciences Analytical Characterization, Merck
This presentation focuses on the utilization of structured data to enable discovery biologics pipeline programs for lead candidate selection and optimization through statistical analysis and machine learning/artificial intelligence (ML/AI) techniques. By leveraging the power of data-driven approaches, this research aims to enhance the effectiveness and efficiency of the lead candidate selection process in drug discovery and development. Through statistical analysis and ML/AI algorithms, the study explores the potential of structured data in identifying and optimizing lead candidates, ultimately facilitating the development of novel and effective therapeutic solutions.
Amber Raines, Senior Director Rapid Analytics, AFS, KBI Biopharma
Flow Imaging is a proven method for characterization of particulates in therapeutic products. Machine learning provides a sophisticated approach to more accurately classify particles in therapeutic products by leveraging the information present in the raw particle images. We will demonstrate how various machine learning algorithms facilitate improved classification compared to the traditional approach, leading to superior sample descriptions.
Brian Berquist, SVP and Chief Development Officer, Wheeler Bio
Wheeler Bio’s Portable CMC™ open-source upstream and downstream platform processes generates a predictable, reliable, and scalable approach for accelerating the movement of molecules from discovery, through lead candidate selection, and into clinical manufacturing. The Portable CMC™ platform and hybrid-mechanistic process model was developed using data from both stable bulk cultures (SBCs, also known as bulk pools) and derivative clones to enable well-controlled cell lines, high titers, process robustness, scalability, and speed-to-clinic.
Networking Coffee Break10:30 am
Standard-Based Strategies for Agile Systems Integration and Interoperability
Serm Kulvatunyou, PhD, Computer Engineer, Process Engineering, National Institute of Standards and Technology (NIST)
Businesses rely on numerous information systems to achieve production goals and improve global competitiveness. Semantically integrating those systems is essential for businesses to achieve both. Using an open data exchange standard for semantic integration has been recognized as a better approach than using a homegrown one. But standard has also been viewed as too slow and oftentimes difficult to use. This presentation will go over recent developments that address these issues. We will also look at recent developments in data standard that may support data exploration and analytics better as well.
SPECIAL PRESENTATION: Ensemble Modeling for the Prediction of Large Molecule Protein Structure
Varsha Daswani, PhD, PMP, Senior Director, Analytics and Data Science, Lumilytics
The in silico methods of identifying three-dimensional protein structure from primary sequence data is an active area of research with applications in drug discovery and development. Large molecule proteins play crucial roles in biological processes and are key targets for drug discovery. However, predicting their complex structures accurately remains a challenging task. We discuss how ensemble modeling, which combines the predictive power of multiple models, can enhance the accuracy and reliability of protein structure predictions in the absence of experimental results.
Session Break12:00 pm
Luncheon Presentation (Sponsorship Opportunity Available) or Enjoy Lunch on Your Own12:05 pm
Session Break1:05 pm
Dan (Cassie) Liu, Principal Statistician, Bristol Myers Squibb
An Integrated System for the Full Automation of Analytical Method Performance Monitoring: A Case Study and Lessons Learned
Method performance monitoring (MPM) is an integral part of analytical method lifecycle to ensure that the methods remain suitable for intended purpose. Routine MPM enables reliable understanding of the long-term method performance, proactive identification and control of method performance changes, and meaningful assessment for product quality-related investigations.
A case study will introduce the fully automated potency MPM system that integrates data collection, data management, visualization, analyses, and results-sharing. This digital advancement remarkably boosts efficiency and productivity by eliminating manual MPM work, reducing wet-lab experiments, bringing more method insights, and allowing quicker data-driven decisions.
Cross-Disciplinary Collaboration: Bridging Industry, Academia, and Public Sector for Life Science Automation Advancements
Life science automation promises enhanced solutions and efficiency but faces challenges like high equipment costs and shortage of skilled personnel. Thus, there's a pressing need for interdisciplinary collaboration and education. The Danish Center for Life Science Automation (DALSA) aspires to be one of the leading European facilities addressing these issues. By creating an interdisciplinary space, DALSA hopes to connect experts and sectors, aiming to bridge gaps and facilitate automation advancements.
Integrating Process and Analytical Data: Trends, Opportunities, and Threats
Jared Auclair, PhD, Director, Bioinnovation; Associate Teaching Professor, Chemistry & Chemical Biology, Northeastern University
This presentation explores the critical role of integrating process and analytical data in biotherapeutic development. We'll discuss current trends, opportunities for improved process optimization and product quality, and potential threats associated with data integration. We'll also explore innovative approaches for data harmonization, advanced analytics, and robust data governance frameworks, highlighting how effective integration can streamline development and elevate biotherapeutic companies.
Brett Averso, MSc, CTO, EVQLV
Embeddings are numerical representations of real-world objects that machine learning systems use to understand complex knowledge domains. Embeddings can capture antibody features, enabling predictions of antibody structure and epitope, even in the absence of crystal structure. Through case studies, we will showcase the method's effectiveness, underscoring significant strides in computational immunology, transforming challenges into opportunities with advances in machine learning.
Eli Bixby, Co-Founder, Head of ML, Cradle
Eli Bixby (Co-Founder & Head of ML) will introduce Cradle’s software platform for ML-based protein optimization. He’ll discuss why many zero-shot generative methods struggle to provide value when moving from chip to lab, and describe the work done to investigate this phenomena. He’ll highlight Cradle’s solutions to issues like poor generalization, batch variance, and diverse batch generation, and present diverse case studies which show the promise of Cradle’s active learning approach.
Special Presentation: Implementation of FDA’s Knowledge-Aided Assessment and Structured Application (KASA) System: Perspectives for CMC Analytics
Bazarragchaa Damdinsuren, MD, PhD, Product Quality Team Leader, OBP, U.S. Food and Drug Administration
KASA is FDA’s internal tool to capture and manage information about intrinsic risk and mitigation approaches for product design, manufacturing, and facilities. The CDER/OPQ is developing KASA for the assessment of biological products as an aid to streamline reviews and increase efficiency. This talk will provide a summary of the current progress and future directions of the biologics KASA platform and discuss the incorporation of CMC analytical functions into KASA.
Networking Refreshment Break3:45 pm
Transition to Plenary Keynote Session4:15 pm
Plenary Keynote Introduction
Laszlo G. Radvanyi, PhD, President & Scientific Director, Ontario Institute for Cancer Research
Driving New CAR T Cells
Marcela V. Maus, MD, PhD, Associate Professor, Medicine; Director, Cellular Immunotherapy, Massachusetts General Hospital
We will talk about various roads and challenges in driving new CAR T cells toward the clinic, and learnings from clinical experience.
High-Throughput Discovery of Protein Folding Stability and Dynamics
Gabriel J. Rocklin, PhD, Assistant Professor, Pharmacology, Northwestern University
Every protein has its own conformational energy landscape that governs its folding stability and dynamics. These varied landscapes are rarely predictable in protein engineering but strongly influence function, aggregation, immunogenicity, and more. Our lab develops new large-scale methods to measure stability and dynamics. I will share lessons from stability measurements of >750,000 protein domains and dynamics measurements of >5,000 domains, highlighting the potential to rationally engineer stability and dynamics.
Welcome Reception in the Exhibit Hall with Poster Viewing6:05 pm
Facilitators of Young Scientist Meet Up: IN-PERSON ONLY
Orhi Esarte Palomero, PhD, Postdoctoral Fellow, Pharmacology, Northwestern University
Alexandros Karyolaimos, PhD, Researcher, Department of Biochemistry & Biophysics, Stockholm University
Shakiba Nikfarjam, PhD, Postdoc, Lawrence Livermore National Lab
Network, Inspire Others and Connect
The young scientist meet-up is an opportunity for scientists entering the field to develop connections across institutions, and for established leaders to come build relationships with the next generation of scientists. The meet-up will pave the way for mentorships, professional opportunities, and scientific discovery.
Close of Day7:30 pm
Tuesday, May 14
Registration and Morning Coffee7:30 am
Chairperson's Remarks
Michail Vlysidis, PhD, Senior Engineer, AbbVie
Protein Language Models Enable Prediction of Polyreactivity of Monospecific, Bispecific, and Heavy-Chain-Only Antibodies
Anusha Prakash, Associate Scientist, AbbVie
Early assessment of antibody polyreactivity is essential for mitigating risks. I will present the development of an ensemble model trained in a transfer learning network to predict the outcomes in the baculovirus particle and bovine serum albumin assays. The training was conducted on a large dataset of sequences augmented with experimental conditions, collected through a highly efficient application. The resulting models demonstrated robust performance on different types of antibodies.
Collection and Management of Binding Data to Support Development of AI and ML Training
Wei Wang, PhD, Senior Principal Scientist, Therapeutic Discovery, Amgen, Inc.
Generative biology incorporates cutting edge biology, high throughput automation and AI to speed up drug development. It’s critical to provide meaningful data to feed machine learning models and to predict developability of the therapeutic candidates. To adapt to the demands of larger sample size and shorter turnaround time, we are transforming our SPR binding assays, including automated assay plate preparation and customized data analysis module in GeneData Screener.
Coffee Break in the Exhibit Hall with Poster Viewing9:00 am
Jennifer R. Cochran, PhD, Senior Associate Vice Provost for Research, Macovski Professor of Bioengineering, Stanford University
Laboratory Evolution of Genome Editing Proteins for Precise Gene Correction and Targeted Gene Integration in Mammalian Cells
David R. Liu, PhD, Richard Merkin Professor and Director, Merkin Institute of Transformative Technologies in Healthcare; Core Institute Member and Vice-Chair of the Faculty, Broad Institute; Director, Chemical Biology and Therapeutic Sciences Program; Investigator, Howard Hughes Medical Institute; Thomas Dudley Cabot Professor of the Natural Sciences and Professor of Chemistry and Chemical Biology, Harvard University
In this lecture I will describe the use of protein evolution and protein engineering to develop precision genome editing technologies. These technologies include base editors, prime editors, recombinases, and CRISPR-associated transposases (CASTs). Base editors and prime editors have been used by many laboratories around the world to correct pathogenic mutations, resulting in ex vivo and in vivo one-time treatments that rescue disease phenotypes in many animal models of devastating genetic disorders. At least nine base editing clinical trials have begun, with positive clinical readouts from at least three of these trials, and the first prime editing clinical trial was recently cleared by FDA to begin in the U.S. I will also describe the use of phage-assisted continuous evolution (PACE) to evolve prime editors, recombinases, and CASTs to enable efficient targeted gene-sized integration in human cells, addressing a longstanding challenge in the genome editing field. These engineered and evolved proteins enable precise target gene correction, disruption, or insertion in a wide range of organisms with broad implications for the life sciences and therapeutics.
Celebrating 20 Years in the Exhibit Hall with Poster Viewing11:00 am
Integrating Computation and Wet-Lab Analysis for the Development of Stable Biotherapeutics
Nicholas Michelarakis, PhD, Postdoctoral Research Fellow, Boehringer Ingelheim Pharmaceuticals
For early-stage development, the time for trial formulation 1 (TF1) decision is a major driver. Due to the vast multidimensional experiment space, finding the most promising formulation remains a major challenge up to this day. Here we will present in silico, sequence-based, structure-based, and machine learning (ML)-based, digital formulation development tools. These approaches leverage the powers of high-throughput experimental measurements in combination with molecular modelling tools.
Machine Learning Models for Immunogenicity and Developability Prediction and Design for Development of New Protein Therapeutics
Jyothsna Visweswaraiah, PhD, Director, Biotherapeutics, Drug Creation, Seismic Therapeutic
The accurate identification of B cell epitopes is crucial to biologics development, but traditional methods for epitope identification are time-consuming and resource-intensive. Therefore, reliable in silico approaches for epitope prediction are needed to accelerate drug discovery.
Jana Hersch, PhD, Head of Scientific Engagement, Genedata
Digital transformation drives change in biopharma companies of all sizes. With the increase in lab automation, growing data stores across all modalities including multispecifics, CGTs, and RNA therapeutics are being leveraged for AI/ML approaches. Biopharma and biotech companies need to connect and structure their data and analytics solutions into robust data streams for R&D projects. I’ll discuss how they do this to extract maximum value, reduce effort duplication, and implement AI/ML.
Session Break1:30 pm
Cindy Gerson, Senior Lead Product Manager, Enterprise Informatics, Schrödinger
Biologics discovery teams are in need of an efficient and comprehensive way to capture and analyze immense amounts of data and to make informed decisions across all stages of the discovery process. This presentation will discuss how LiveDesign, Schrödinger’s enterprise informatics platform, has been developed to enable biologics workflows. LiveDesign expedites and improves decision making by 1) uniting all in vitro and in silico data and metadata in one centralized hub, 2) supplying in-platform tools for decision tracking and streamlined communication, 3) integrating and democratizing computational modeling execution, delivery, and 3D visualization, and 4) providing an agnostic snap-in/snap-out framework for flexible evolution of workflows.
Luncheon Presentation (Sponsorship Opportunity Available) or Enjoy Lunch on Your Own2:10 pm
Close of Digital Integration in Biotherapeutic Analytics Conference2:40 pm
Recommended Dinner Short Course6:30 pm
SC8: Developability of Bispecific Antibodies
View By:
May 12-13
Display of Biologics
Antibodies for Cancer Therapy
Advances in Immunotherapy
Difficult-to-Express Proteins
ML and Digital Integration in Biotherapeutic Analytics
Biologics for Immunology Indications
May 13-14
Engineering Antibodies
Advancing Multispecific Antibodies
Emerging Targets for Oncology and Beyond
Engineering Cell Therapies
Optimizing Protein Expression
Biophysical Methods
Predicting Immunogenicity with AI/ML Tools
Radiopharmaceutical Therapies
May 15-16
Machine Learning for Protein Engineering
Driving Clinical Success in Antibody-Drug Conjugates
Engineering Bispecific and Multifunctional Antibodies
Next-Generation Immunotherapies
Maximizing Protein Production Workflows
Characterization for Novel Biotherapeutics