complexa-design

Automate protein binder design by orchestrating the full generative lifecycle, including initial generation, filtering, evaluation, and success-rate analysis, to reduce manual configuration. It handles the entire workflow from target input to final manifest generation.

Generative biomolecular modeling for pocket scaffolding designs novel ligand-binding sites and enzyme active sites. It automates the multi-stage scientific workflow to generate and evaluate structural scaffolds, preventing common manual configuration errors.

Yes, generative chemistry workflows require a configured environment with CUDA-enabled GPU resources and specific model checkpoints for AF2, RF3, and MPNN backends to execute the folding and reward evaluations.

Yes, you can use diverse folding backends including AF2, RF3, and ESMFold for flexible protein design iteration. The pipeline integrates these backends to manage the generation and evaluation lifecycle.

Designing a de novo binder for targets like PDL1 is best handled by running a fully orchestrated pipeline using default beam search configuration. It processes the target through generation, filtering, and evaluation automatically.