Life Science

Accelerating Drug Discovery with Advanced Computing and AI: Inside the Schrödinger – AstraZeneca Collaboration

In late 2019, Schrödinger and AstraZeneca teamed up with the aim of accelerating the discovery of novel therapeutic compounds by improving the throughput of binding affinity predictions. After a successful pilot study, the technology and workflows are now rolled out across AstraZeneca’s small molecule discovery programs.

By augmenting traditional drug design with sophisticated computational methods to predict what molecules to make next and how to make them, this partnership will ultimately enable us to bring better medicines to patients faster.

Read on to hear our perspectives on this new approach to drug discovery:

Anna Åsberg, Vice President, R&D IT, AstraZeneca

Making, testing and analyzing compounds is both resource- and time-intensive. At AstraZeneca, we have made a decision to put digital technology at the heart of our strategy to accelerate this process.

Schrödinger has developed a computational platform to integrate active learning with their physics-based predictive modeling tools such as FEP. This enables the rapid optimization of molecules with regards to binding affinity – an important factor in the design of new compounds – and was an attractive approach to include into our drug discovery projects.

In 2020, we ran a pilot study on more than 10 small molecule projects. We demonstrated a 10-100 fold increase of throughput in the number of FEP calculations as compared to what we could run without combining active learning into FEP (AL-FEP). This increase allowed us to assess and evaluate potential compounds far more accurately and quickly, reducing the number of compounds synthesized in the laboratory.

Werngard Czechtizky, Head of Medicinal Chemistry, Research and Early Development, Respiratory and Immunology, AstraZeneca

For our pilot program, we worked on a variety of different targets. The targets were structurally enabled – a prerequisite for using Schrödinger’s AL-FEP platform effectively. The designs resulting from the Schrödinger platform calculations are used to guide which compounds we synthesize for each project, which saves time and money.

We identified improved and/or novel ligands for half of the targets, including some that traditionally have been difficult to drug. Right now, we are already moving several of these projects through preclinical milestones.

The promising results have led us to extend the collaboration by another three years to fully harness the AL-FEP platform within our computational design. We now aim to use this approach to drive value across all small molecule portfolio projects that are based on structurally-enabled targets.

Robert Abel, Ph.D., Executive Vice President and Chief Computational Scientist, Schrödinger

This collaboration is another important proof point demonstrating that our platform truly has the potential to transform the discovery of new medicines.

We’ve seen the impact our platform can have on our own internal pipeline, in the biotech companies we have co-founded and in our collaborations with global pharmas around specific targets and indications. Now, we are seeing it deployed on a larger scale by a global leader in integrating cutting-edge digital technology into drug discovery.

The success of this collaboration is very gratifying – but it’s important to note that we’re not stopping here. Continued innovation is woven into our DNA at Schrödinger. We invest heavily in R&D to improve and refine our platform and to extend its utility to new targets. To that end, we established an additional collaboration with AstraZeneca in March 2020 to refine our computational platform for biologics. We’re hopeful that we can match the success we’ve shown in the small molecule realm into the design of novel antibodies and protein-based therapeutics.

This is an exciting time for computation-driven drug discovery. We’re thrilled to be working with AstraZeneca, an excellent partner in the quest to accelerate and improve the search for novel therapies, and we look forward to seeing what else we can achieve together.

Anna Åsberg

Anna Åsberg is vice president of R&D IT, responsible for strategy, build and run for IT in both Oncology and BioPharmaceuticals R&D. Her global team delivers transactional core IT systems as well as transformational data and insight technologies, including engineering, platform and tools for artificial intelligence and digital health capabilities. Anna holds a Master of Science degree from KTH, Royal Institute of Stockholm. Prior to joining AstraZeneca in 2004, Anna studied and worked in numerous European countries, including Germany, Spain, Switzerland and France. She developed a passion for applying IT to diverse organizations and challenges, and launched an IT consultancy company in Hamburg, Germany. Her AstraZeneca career began in Södertälje, Sweden, working with Finance and Enabling Units, followed by a move to the US to provide IT for the US Commercial unit. Anna moved back to Södertälje in 2014 as VP for Operations IT, where she established and started the journey to make AstraZeneca factories ‘smart’ with digital technologies.

Werngard Czechtizky

Werngard Czechtizky is Chair of AstraZeneca’s Global Chemistry Leadership and Head of Medicinal Chemistry, Early Research and Development, Respiratory and Immunology, BioPharmaceuticals R&D. She has a track record of delivering clinical candidates and lead compounds across several therapeutic areas including CV, Diabetes, Pain, CNS, Inflammation, Respiratory and Immunology.
Werngard studied at the Technical University of Graz, Austria. After working at Bayer in Pittsburgh, USA, she received a Ph.D. from ETH Zürich and a postdoctoral training at Harvard University. She worked at Aventis, then Sanofi in Germany, before joining AstraZeneca in 2017.

Robert Abel

Robert Abel joined Schrödinger in 2009 and is responsible for advancing Schrödinger’s computational science platform. He also leads the computational chemistry team within Schrödinger’s drug discovery group. Robert obtained his Ph.D. from Columbia University, where he was awarded NSF and DHS research fellowships. His thesis work with Professor Richard Friesner involved developing methods to quantify the role of solvent in protein-ligand binding. Robert has co-authored multiple patent applications and continues to publish extensively on a wide variety of topics in computational chemistry.

Sign Up Today


Sign up to receive quarterly updates with the latest from Extrapolations.