Search for: All records

A popular approach to protein design is to combine a generative model with a discriminative model for conditional sampling. The generative model samples plausible sequences while the discriminative model guides a search for sequences with high fitness. Given its broad success in conditional sampling, classifier-guided diffusion modeling is a promising foundation for protein design, leading many to develop guided diffusion models for structure with inverse folding to recover sequences. In this work, we propose diffusioN Optimized Sampling (NOS), a guidance method for discrete diffusion models that follows gradients in the hidden states of the denoising network. NOS makes it possible to perform design directly in sequence space, circumventing significant limitations of structure-based methods, including scarce data and challenging inverse design. Moreover, we use NOS to generalize LaMBO, a Bayesian optimization procedure for sequence design that facilitates multiple objectives and edit-based constraints. The resulting method, LaMBO-2, enables discrete diffusions and stronger performance with limited edits through a novel application of saliency maps. We apply LaMBO-2 to a real-world protein design task, optimizing antibodies for higher expression yield and binding affinity to several therapeutic targets under locality and developability constraints, attaining a 99% expression rate and 40% binding rate in exploratory in vitro experiments. 

Structure-aware machine learning captures strain impact on molecule-surface interactions for rapid catalyst evaluation. 

Related Article: Diane A. Dickie, Nathan C. Frey and Robert J. Gilliard Jr.|2022|CSD Communication||| 

The correlation consistent Composite Approach for transition metals (ccCA-TM) and density functional theory (DFT) computations have been applied to investigate the fluxional mechanisms of cyclooctatetraene tricarbonyl chromium ((COT)Cr(CO)3) and 1,3,5,7-tetramethylcyclooctatetraene tricarbonyl chromium, molybdenum, and tungsten ((TMCOT)M(CO)3 (M = Cr, Mo, and W)) complexes. The geometries of (COT)Cr(CO)3 were fully characterized with the PBEPBE, PBE0, B3LYP, and B97-1 functionals with various basis set/ECP combinations, while all investigated (TMCOT)M(CO)3 complexes were fully characterized with the PBEPBE, PBE0, and B3LYP methods. The energetics of the fluxional dynamics of (COT)Cr(CO)3 were examined using the correlation consistent Composite Approach for transition metals (ccCA-TM) to provide reliable energy benchmarks for corresponding DFT results. The PBE0/BS1 results are in semiquantitative agreement with the ccCA-TM results. Various transition states were identified for the fluxional processes of (COT)Cr(CO)3. The PBEPBE/BS1 energetics indicate that the 1,2-shift is the lowest energy fluxional process, while the B3LYP/BS1 energetics (where BS1 = H, C, O: 6-31G(d′); M: mod-LANL2DZ(f)-ECP) indicate the 1,3-shift having a lower electronic energy of activation than the 1,2-shift by 2.9 kcal mol−1. Notably, PBE0/BS1 describes the (CO)3 rotation to be the lowest energy process, followed by the 1,3-shift. Six transition states have been identified in the fluxional processes of each of the (TMCOT)M(CO)3 complexes (except for (TMCOT)W(CO)3), two of which are 1,2-shift transition states. The lowest-energy fluxional process of each (TMCOT)M(CO)3 complex (computed with the PBE0 functional) has a ΔG‡ of 12.6, 12.8, and 13.2 kcal mol−1 for Cr, Mo, and W complexes, respectively. Good agreement was observed between the experimental and computed 1H-NMR and 13C-NMR chemical shifts for (TMCOT)Cr(CO)3 and (TMCOT)Mo(CO)3 at three different temperature regimes, with coalescence of chemically equivalent groups at higher temperatures.