Merck announced on June 17 a new collaboration with California-based Protillion Biosciences to integrate large-scale data generation and artificial intelligence design for the advancement of novel biologic therapies.
According to the announcement, Merck will make an undisclosed upfront payment and could pay up to $510 million in research, development, and commercial milestone payments. The companies did not disclose specific targets or disease areas but stated they would work together on multiple therapies.
As part of the agreement, Merck will gain access to Protillion’s Prot-MaP technology. This platform is described as a "megascale data generation platform" that provides training sets specifically tailored for protein design AI models. According to information from Protillion’s website included in the release, this approach allows simultaneous and quantitative assessment of protein therapy candidates at a large scale. The goal is to identify candidates likely to be specific for their intended targets while maintaining manufacturability.
Juan Alvarez, vice president of discovery biologics at Merck Research Laboratories, said in a statement that Protillion’s technology "offers a compelling opportunity" for Merck’s drug development capabilities. He added that the platform has the "potential to transform the speed and precision with which we characterize protein landscapes and identify novel therapeutic candidates."
The announcement comes as artificial intelligence becomes increasingly central within biopharma drug development processes. Recent industry partnerships include Eli Lilly investing up to $2.25 billion with Profluent Bio for machine learning-driven enzyme design, as well as alliances by Bristol Myers Squibb and Incyte focused on integrating machine learning models into various operations such as data sharing and commercial functions. Earlier this month, Alnylam committed $30 million upfront—and up to $2 billion total—to partner with Inceptive Nucleics on optimizing siRNA design using machine learning engines.
