binder-design

To plan a protein binder design campaign, prepare your target structure, select interface hotspots, run BoltzGen to generate candidate binders, validate structures with Chai, and apply QC filters for pLDDT, ipTM, and PAE_interface before experimental testing.

Choosing between BoltzGen and BindCraft depends on your target type and design goals; this workflow provides a decision tree to select the optimal tool, guiding whether to use BoltzGen, BindCraft, IgGM, or mBER for your specific binder campaign.

To validate protein binders, run generated structures through Chai validation and apply QC filtering thresholds for pLDDT, ipTM, and PAE_interface scores to ensure structural confidence and interface quality before proceeding to experimental testing.

Yes, you can use this workflow for peptide, antibody, and nanobody binder design scenarios. It accommodates various target contexts including ligand-context targets, helping you specify target preparation and hotspot selection for these specific binder types.

Yes, you need GPU-backed model execution resources to run the recommended protein binder design pipeline. Runtime resource expectations include GPU support for executing the BoltzGen generation and Chai validation models effectively.

De novo protein binder designs often fail validation due to poorly constrained targets or inadequate hotspot selection. Specifying proper target preparation, accurate hotspot selection, and strict QC filters for pLDDT, ipTM, and PAE_interface prevents wasted compute and validation bottlenecks.