Biomolecular interactions, including protein–protein interactions, protein–nucleic acid recognition, and protein–small molecule binding, underlie a wide range of biological processes and therapeutic mechanisms. Although recent de novo design methods can generate candidate binders for diverse molecular targets, practical design campaigns remain limited by low filter-passing rates and model-specific biases that arise when designs are optimized against a single predictor. Here, we present RFOptimization (RFO), a training-free framework for all-atom biomolecular binder optimization. RFO formulates binder improvement as a residue-wise mutational search problem, sampling candidate substitutions alternately based on gradient-guided sequence optimization using all-atom structure prediction models and a cycling-based sequence redesign strategy that alternates structure generation with an orthogonal predictor and MPNN-based sequence design to improve the in silico success rate of RFdiffusion-generated binders within minutes of computation. To reduce overfitting to any individual structure model, candidate mutations are further evaluated with orthogonal AlphaFold3 metrics as final filters. We demonstrate the generality of RFO across diverse design settings, including classical protein binder design, ligand-binding biosensor design, cyclic peptide design, and active site-aware enzyme design.
Odin Zhang, Jia-Qi Wang, T. Thompson et al.· bioRxiv· 0 citations
Acyl-homoserine lactone (AHL) quorum sensing enables many species of proteobacteria to coordinate collective behaviors. In such systems, a synthase produces an AHL signal, which is sensed by a LuxR-family receptor. Despite extensive genomic annotation of LuxR homologs, preferred AHLs for most receptors remain unknown, limiting functional understanding of quorum sensing across diverse bacteria. Here, we present the COPAL (combining ordered predictions of audited ligands) pipeline, which integrates multiple protein-ligand co-folding models to identify preferred AHLs for a specific LuxR. Benchmarking on a leakage-controlled subset of 96 experimentally characterized LuxR-AHL pairs shows that COPAL places the preferred AHL within the top-6 candidates (out of 58) for 68% of receptors, outperforming every individual co-folding model. Further, inter-model agreement correlates with ranking accuracy, offering an indication of confidence. We show that COPAL resolves the specificity shift induced by three-point mutations in LasR and correctly nominates C8-HSL as the preferred ligand for the previously uncharacterized Mesorhizobium sp. NJ3 receptor, which we verified experimentally. Finally, we release the Ranked AHL-LuxR Prediction Hub (RALPH), comprising precomputed rankings for about 10,000 unique LuxR homologs. More broadly, COPAL shows that unweighted rank aggregation of complementary co-folding models offers a general strategy for predicting receptor-ligand specificity in data-scarce biological systems.
Davi Nakajima An, A. Schaefer, Szu-Min Chang et al.· bioRxiv· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
