ACS Spring 2025
Date & Time - March 23rd-27th, 2025
Location - San Diego, California
Schrödinger is excited to be participating in the ACS Spring 2025 conference taking place on March 23rd – 27th in San Diego California. Stop by booth #3532 to speak with Schrödinger scientists or join us for happy hour on March 26th. Full event details below.
MAR 23 | 8:00AM 
Ballroom 20D Unlocking a New Era: How AI is Transforming Drug Discovery and Development
Speaker:
Yefen Zou, Senior Principal Scientist, Schrödinger
Abstract:
We are seeing an unprecedented growth on AI-driven research towards the development of new medicines. Nowadays, AI can impact all aspects of drug development: from enabling generative chemistry, to analysis of large datasets to accelerate hit finding, reduce timelines in the design of drug-like lead compounds, to structure-based drug design with AlphaFold to applications to green and sustainable drug development synthesis. This exponential growth covers AI-biotechs to large pharma, pioneer academic researchers and AI-technology organizations, effectively unlocking a new era to bring new medicines to patients in need. This symposium will showcase examples of how AI is impacting drug development from early discovery, to drug development, from industry to academic research and teams comprising members of MEDI, ORGN, COMP, I&EC, CINF Divisions.
MAR 23 | 3:45PM
Pacific Ballroom: Section 23 Transforming Polymer Design for Industrial Applications Using Experiment Data, Machine Learning, and Physics-Based Modeling
Speaker:
Atif Afzal, Principal Scientist II, Schrödinger
Abstract:
Designing industrially relevant polymers is challenging due to the need to simultaneously optimize multiple material properties, making traditional trial-and-error methods costly and inefficient. A promising alternative is the integration of machine learning (ML) and physics-based approaches to explore the polymer design space and identify candidates that meet specific industrial criteria. This work presents a workflow combining ML and molecular modeling techniques, demonstrated through two case studies in polymer design.
The first case is an internal study focused on elastomers with target thermophysical and mechanical properties. Using curated datasets of experimental property values from the literature, we developed ML models that accurately predict the properties of new elastomers. These models screened over 100,000 copolymers, identifying top candidates with superior glass transition temperatures and elastic moduli suitable for elastomer applications.
The second case presents results from a collaboration with SABIC Specialties involving polycarbonate design*, targeting key polymer properties, including glass transition temperature, optical properties, and mechanical strength. Using an experimental dataset, we trained ML models to predict these polymer properties based on structure and monomer composition. We subsequently screened several thousands of polycarbonate structures generated via R-group enumeration and identified the top candidates. These candidates were validated through molecular dynamics simulations and density functional theory, demonstrating strong correlations between ML predictions and physics-based approaches. Experimental validation further confirmed the accuracy of our models.
This workflow demonstrates the power of integrating data-driven and physics-based methods in polymer design, offering an efficient strategy for material scientists working with limited experimental data.
*Collaboration with SABIC Specialties is gratefully acknowledged.
MAR 24 | 11:35AM Combined Physics-Based and Machine Learning Approaches in the Design of Complex Materials
Speaker:
Anand Chandrasekaran, Senior Principal Scientist, Schrödinger
Abstract:
The simulation of materials properties using physics-based approaches, such as density functional theory (DFT) and molecular dynamics (MD), has long been successful in providing insights into structure-property relationships and subsequently aiding in the design of novel materials. More recently, machine learning (ML) has been used extensively in conjunction with physics-based modeling techniques to greatly accelerate materials innovation. The accuracy and generalizability of physics-based modeling improves the performance of ML models and enables them to be used effectively even in small-data regimes. Conversely, the speed and flexibility of ML help bridge the time- and spatial- scale limitations of physics-based models, creating a synergistic approach that optimizes both predictive accuracy and computational efficiency. In this talk, we demonstrate the application of this combined approach in designing materials and formulations across diverse applications, from battery electrolytes and fuel mixtures to thermoplastics and OLED devices. For instance, we demonstrate how DFT descriptors greatly improve the accuracy of ML models for optoelectronic molecules and battery electrolytes while descriptors from MD simulations can lead to better models for viscosity of organic molecules. We use Schrodinger’s automated Formulation ML solution, which takes into account both chemistry and composition, to train ML models for solubility of APIs in binary solvents and for the prediction of motor octane number of hydrocarbons. Additionally, we showcase recent advancements in our machine learning force field technology (MPQRNN), which has been trained on a vast chemical space encompassing over 86 elements, and demonstrate its application in accurately modeling the bulk properties of inorganic cathode coating materials.
MAR 25 | 10:30AM
Room 25C Leveraging Cloud Computing to Efficiently Identify the Most Promising Compounds in Ultra-Large Chemical Spaces for In-Silico Hit Discovery
Speaker:
Steven Jerome, Executive Director, Schrödinger
Abstract:
By screening ultra-large libraries in the cloud with a per-target tailored, hierarchical screening approach, the dream of achieving double-digit hit rates and diverse starting chemical matter in virtual screening has been achieved for multiple drug discovery projects. Virtual libraries for in-silico hit finding can be thought of as curated subsets of larger chemical spaces defined by a set of reactions and matching reagent libraries. While many commercial vendors provide such “ready to screen” virtual libraries, advances in cloud computing make it possible to tailor custom libraries on a per-project basis by exploring the full vendor space. At Schrödinger, project teams begin all virtual screening campaigns by searching an internal cloud database built on Google’s BigQuery comprising more than 145 billion purchasable compounds representing 50 vendors, including fully enumerated ultra-large vendor spaces such as Enamine Real. In order to identify the most promising compounds for virtual screening, we have developed a screening methodology called QuickShape based on a custom 1D fingerprint which aims to capture pharmacophore-like features. This compact representation is well-suited for representing ultra-large chemical spaces and is incorporated into our cloud database. In this talk, we present our cloud-native approach for the generation of target-specific libraries for virtual screening together with a pair of prospective studies inside active drug discovery programs, covering both fragment and druglike molecule virtual screening.
MAR 25 | 12:00PM
Room 1A 20+ years of AI in Drug Discovery: From Promise to Impact
Panelist:
H. Rachel Lagiakos, Director, Medicinal Chemistry, Schrödinger
Abstract:
Talk followed by a panel discussion.
MAR 26 | 9:00AM
Room 28A/B MEDI First Time Disclosures
Host:
H. Rachel Lagiakos, Director, Medicinal Chemistry, Schrödinger
Abstract:
The highly anticipated session that chronicles the journey of a molecule from its discovery on the bench to its progression into clinical trials. This session emphasizes the challenges and successes that medicinal chemists face every day, and reminds us of the tremendous impact our efforts can bring!
MAR 26 | 6:00PM MEDI First Time Disclosures Networking Reception
Hosted by Schrödinger
Join other medicinal chemists at this happy hour event for networking and drinks. Hosted outside room 28A/B in the San Diego Convention Center
