A new multimodal tool combines a large language model with powerful graph-based AI models to efficiently find new, synthesizable molecules with desired properties, based on a user’s queries in plain language.

Investigators from Mass General Brigham and Beth Israel Deaconess Medical Center have developed STITCHR, a new gene editing tool that can insert therapeutic genes into specific locations without causing unwanted mutations. The system can be formulated completely as RNA, dramatically simplifying delivery logistics compared to traditional systems that use both RNA and DNA. By inserting an entire gene, the tool offers a one-and-done approach that overcomes hurdles from CRISPR gene editing technology—which is programmed to correct individual mutations—offering a promising step forward for gene therapy. Results are published in Nature.