Cytobank, FDA using machine learning to classify MSCs

Cytobank Inc. (Santa Clara, Calif.) partnered with FDA to develop machine learning-based methods for characterization of mesenchymal stromal cell morphology that could be used to improve cellular manufacturing.

The collaboration is based on experiments by Steven Bauer, chief of the cellular and tissue therapy branch at FDA’s CBER, which identified MSC morphological features that were predictive of immunosuppressive capacity and levels of interferon (IFN) lambda-mediated immunosuppression enhancement.

Cytobank will apply its cloud-based informatics platform to imaging and phenotype data from FDA to develop processes for classifying cellular preparations according to their immunotherapeutic properties.

Cytobank President and CEO David Craford told BioCentury IP from the deal will be shared by the two organizations. Craford said the partnership is a cooperative agreement with no money changing hands.