BenchSci and Thermo Fisher Collaborate to Develop AI Tools for Preclinical R&D

Toronto-based BenchSci has entered a strategic partnership with Thermo Fisher Scientific to create AI-powered tools aimed at improving experimental design, reagent selection, and overall R&D productivity. The collaboration combines BenchSci’s ASCEND AI platform with US-based Thermo Fisher’s laboratory instruments, consumables, and services to support preclinical and translational research.

#advertisement

The companies plan to develop enterprise software for scientific instruments, web-based research literature search, and optimized reagent workflows. These tools are designed to help researchers interpret large datasets, refine experimental strategies, and make better use of laboratory resources. BenchSci’s approach is built on a proprietary biomedical knowledge graph and machine learning models that process published data to extract biological insights.

ASCEND is BenchSci’s platform built to help researchers navigate the growing overload of biomedical information. Each year, more than a million new scientific papers are published and vast amounts of experimental data are generated, much of it scattered across inconsistent formats. Traditional language models can summarize this material but often fail in scientific settings because they cannot reliably distinguish ambiguous terms or trace claims back to the original evidence. 

See also: Google-Backed BenchSci Reports New System to Extract Experimental Context From Biomedical Papers Into a Traceable Evidence Graph

ASCEND addresses these issues by combining generative AI with a large biomedical knowledge graph that encodes over 400 million entities and one billion relationships, linking genes, proteins, diseases, reagents, and outcomes directly to experimental sources such as publications, patents, and reagent catalogs. 

The result is a system where researchers can ask questions in natural language and receive answers grounded in verifiable data, supporting reproducibility and reducing duplication of effort. Our recent case study illustrates how ASCEND structures this fragmented knowledge into a machine-readable map for biomedical reasoning.

Topic: AI in Bio