As high-resolution cryo-EM becomes routine for biologics characterization, structure-guided protein design/engineering, increasingly augmented by machine learning, is transitioning from a specialized capability to a core accelerator of R&D timelines. However, many teams continue to encounter common bottlenecks: construct design decisions made too late in the process, inconsistent solubility/stability outcomes, and limited ability to engineer specific properties reliably without repeated iteration cycles.

This practical workshop will equip attendees with a structure-first, ML-informed framework for biologics engineering, emphasizing developability as critical gatekeepers for downstream success. Using literature-based examples and widely accessible tools such as Pymol, the session will demonstrate how structural biologists assess targets, select constructs, and propose engineering strategies achievable within industrial timelines.

Workshop Highlights:

• Construct Selection Strategies: Practical approaches to selecting constructs early in the design process including domains, truncations, linkers, and formats to enhance expression and structural tractability
• Integrated Feedback Loops: How experimental, data science, and ML teams collaborate to transform structural and biophysical data into fewer design iterations
• Hands-On Demonstrations: Step-by-step walkthroughs using published structures and common tools, complemented by peer discussion on how attendees approach similar decisions in their own work