Machine learning-guided design of mechanoadaptive bioglues for multitissue trauma and first-aid applications

Treating wounds that involve multiple types of injury is particularly challenging due to their complex morphology and the diverse mechanical properties of the affected biological tissues. Here we report a machine learning (ML)-guided rational design of bioglues, termed TuneGlues, tailored with mechanical adaptability for multitissue trauma. By leveraging ML, we establish task-oriented relationships between TuneGlues and various tissues, enabling precise optimization for specific targets. Four representative TuneGlues, which are developed and tested for lung, intestine, skin and bone injuries, demonstrate promising adhesive properties and postoperative healing outcomes. In addition, we integrate a comprehensive mechanical database derived from the ML model into a custom first-aid device capable of rapidly delivering optimized TuneGlues to target tissues. This system notably reduces the duration of wound treatment and enhances outcomes for multitissue trauma in open surgeries. The integration of ML-guided TuneGlues design with a first-aid delivery device provides a transformative strategy for emergency care, advancing the field of multitissue trauma treatment and tissue engineering.