Blog Post by Reveal Bio follow-up to How is AI Benefiting Biopharma?

BIOCOM CRO EVENT: How is AI Benefiting Biopharma?

This Biocom event featured a full house with lively discussion as we learned about how Reveal Biosciences and Immuneering Corporation are applying Artificial Intelligence (AI) to benefit biopharma. Our speakers were:

Casey Laris, CTO
Reveal Biosciences

Rebecca Kusko, PhD, SVP & Head of R&D
Immuneering Corporation

Casey began the program with an overview of AI, describing a rapid growth in investment in this sector as well as a corresponding decrease in the cost of compute resources that accelerates analysis to a commercially viable rate. Reveal specializes in using AI to analyze pathology whole slide images with applications in preclinical research, clinical trials, and diagnostics. Each whole slide image is approximately 1 – 3 GB of data, and the company is analyzing many thousands of images per year.

Reveal has developed an AI training, inference and visualization platform called ImageDx™ for pathology images. They are using this platform for disease specific applications within therapeutic areas including oncology, NASH, and dermatology. Within biopharma these applications can be used to increase the accuracy, reproducibility and scale of pathology analysis. The data can also be combined with other data forms such as clinical outcome, biomarker data, or RNA-Seq to build predictive AI-based models for patient stratification.

Immuneering’s drug discovery platform leverages advanced computational systems to reduce R&D costs and improve the translational success of novel drugs. Their focus to date has been in CNS, inflammatory disease and oncology. Rebecca showed compelling data from two drug discovery platform modules: Cosiner and Fluency. Cosiner is a hypothesis free application for target identification and in silico screening. This module uses input data including disease biology and compound gene expression to generate disease profiles in weeks with compound screening completed in just hours.

The second module, Fluency, applies AI-based prediction of protein-drug binding interactions to screen additional compounds against identified targets. The input data for this module is protein amino acid sequences and small molecule SMILES strings. Application of this module can predict the IC50 for a protein across 13M compounds in only 5 hours. Taken together, these Immuneering in silico screening modules can result in significant time and cost savings for biophama companies.

In summary, AI is being used increasingly by biopharma companies to accelerate workflow and reduce costs for specific applications. This powerful technology can also generate data to stratify patient groups or reduce drug toxicity in ways that were previously not possible. As the industry increasingly focuses on precision medicine, AI-based tools such as those developed by Reveal and Immuneering address a strong need in the market.



About the Author and BIOCOM CRO Board Member