Bayesian Optimization for Ternary Complex Prediction (BOTCP)

Abstract

Proximity-inducing compounds (PICs) are an emergent drug technology through which a protein of interest (POI), often a drug target, is brought into the vicinity of a second protein which modifies the POI's function, abundance or localisation, giving rise to a therapeutic effect. One of the best-known examples for such compounds are heterobifunctional molecules known as proteolysis targeting chimeras (PROTACs). PROTACs reduce the abundance of the target protein by establishing proximity to an E3 ligase which targets the protein towards degradation via the ubiquitin-proteasomal pathway. Design of PROTACs in silico requires the computational prediction of the ternary complex consisting of POI, PROTAC molecule, and the E3 ligase. Here, we present a novel machine learning-based method for predicting PROTAC-mediated ternary complex structures using Bayesian optimization. We show how a fitness score combining an estimation of protein-protein interactions with PROTAC binding energy calculations enables the sample-efficient exploration of candidate structures. Furthermore, our method presents two novel scores for filtering and reranking which take PROTAC stability (Autodock-Vina based PROTAC stability score) and protein interaction restraints (the TCP-AIR score) into account. We evaluate our method using DockQ scores and demonstrate, that even with a clustering that require members to have a high similarity, i.e. with smaller clusters, we can assign high ranks to those clusters that contain poses close to the experimentally determined native structure of the ternary complexes. We also demonstrate the resultant improved yeild of near-native poses in these clusters.

Advancing Targeted Protein Degradation via Multiomics Profiling and Artificial Intelligence

“…71 In this line, efficient optimization techniques have been proposed to sample the space of ternary complex candidates using composite scores that take into account both PPI docking results and a PROTAC score that measures the feasibility of accommodating the ligand given the position of the binding pockets. 250 Additional biological constraints can be incorporated to further restrict the search space, such as the location of the lysine in the POI for ubiquitination, or of known residue mutations that modulate complex formation. Overall, though, given the abundance of rotatable bonds in the small-molecule degrader, and the importance of protein backbone flexibility in the coordinated formation of PROTAC ternary complexes, modeling these assemblies is likely to remain a computationally intensive task despite the constraints imposed by PROTACbridged complexes.…”

“…Similarly, the computational pipeline PRosettaC starts by modeling POI-warhead and E3-recruitment ligand interactions, and then exploits the fact that PROTAC linkers have length constraints to drastically reduce the docking search space (Figure ). In this line, efficient optimization techniques have been proposed to sample the space of ternary complex candidates using composite scores that take into account both PPI docking results and a PROTAC score that measures the feasibility of accommodating the ligand given the position of the binding pockets . Additional biological constraints can be incorporated to further restrict the search space, such as the location of the lysine in the POI for ubiquitination, or of known residue mutations that modulate complex formation.…”

Targeting the Epidermal Growth Factor Receptor with Molecular Degraders: State-of-the-Art and Future Opportunities

“…In a similar context, another pharmaceutical giant, Celeris Therapeutics in Graz, Austria, has developed an in silico degrader design (BOTCP) to design E3 ligases. 216 This allows the design via an agnostic method to develop small molecules with improved physiochemical assets. This design is also vital as it ensures the possibility of ternary complexes formation in a very short span of time with unprecedented accuracy.…”

“…Zeptomics has been explored to identify ligands for both POI and degraders, design heterobifunctional proteins, and prediction of safety and efficacy of the designed PROTACs. In a similar context, another pharmaceutical giant, Celeris Therapeutics in Graz, Austria, has developed an in silico degrader design (BOTCP) to design E3 ligases . This allows the design via an agnostic method to develop small molecules with improved physiochemical assets.…”

Current Challenges in Small Molecule Proximity-Inducing Compound Development for Targeted Protein Degradation Using the Ubiquitin Proteasomal System

“…Next to these approaches based on the small molecule itself, especially in targeted protein degradation, several solutions have been developed to tackle the identification of ligandable pockets [ 43 ], modelling of ternary structures [ 44 ], prediction of degradation [ 45 ] and also general prediction of protein structures in the absence of structural data [ 46 ]. Moreover, methods developed for docking and virtual screening [ 47 , 48 ] are being employed for TPD warhead and recruiter identification.…”