By Jane Cook
November 5, 2021
Machine Learning Saves Time
What is the biggest barrier to optimizing the cancer therapy development pipeline? The simple answer is time. Each step in the pipeline can take a year or more from designing a drug structure to testing drug candidates and looking for synergistic effects with existing treatments.
Now, however, bioinformatics and machine learning are shrinking the time it takes to execute several stages of the cancer drug discovery process. Machine learning uses algorithms and models to make predictions from data, with the ability to “learn” how to make these predictions over time without explicit instructions.
Effective Deep Learning Models
A 2019 paper in Nature Biotechnology developed a deep generative model for the de-novo design of small molecules using reinforcement learning called GENTRL. Using GENTRL, the researchers discovered a structure for a potent inhibitor of DDR1, a kinase involved in the progression of numerous cancers. The traditional timeline for the successful discovery of that nature is a year or more, while GENTRL identified the new drug in just 21 days.
Machine Learning and Predictive Models for Combination Therapy
Another advantage of using machine learning is being able to model drug interactions for potential toxicity and efficacy. Combination therapy, or the use of multiple drugs simultaneously, is a popular approach in cancer therapy but is time intensive and difficult to test due to off-target effects, unforeseen toxicity, and highly variable patient responses.
Machine learning algorithms help build predictive models. And those models can optimize combination therapy development by identifying the individual targets present in a patient’s tumor and the most effective combination of drugs to reach those targets with minimal toxicity.
Personalized Medicine: A New Approach
This is a true personalized medicine approach, and a machine learning algorithm developed in 2018, CURATE.AI, helps design combination therapy using only a single patient’s data. A patient treated using the designed therapy from CURATE.AI had their tumor progression stopped and exhibited a durable response to the treatment. This success has led to this machine learning-based approach being tested in a clinical trial.
Machine learning and bioinformatics are already significantly improving the efficiency of the cancer drug discovery and development pipelines, and it is still just the beginning of this new approach to therapy development.
Jane Cook, Journalist & Content Writer, Bridge Informatics
Jane is a Content Writer at Bridge Informatics, a professional services firm that helps biotech customers implement advanced techniques in the management and analysis of genomic data. Bridge Informatics focuses on data mining, machine learning, and various bioinformatic techniques to discover biomarkers and companion diagnostics. If you’re interested in reaching out, please email [email protected] or [email protected].