Thesis advisor: Junwei Bao / The mysteries in chemistry could be reveals by utilization of innovative computational methods and the creative application of advanced modeling techniques. The limitation or difficulties of experimental methods sometimes makes it hard to study the mechanism behind the reaction energy, so computational chemistry could play an important role to investigate and analyze the detailed reaction mechanism and elementary reaction pathways behind a complexed, multi-step chemical reaction. Also, it is important to develop robust and accurate theoretical methods to perform computational and simulation works. We applied multiple computational or simulation models to study various chemical systems, which provided us valuable insights to understand the chemical reactions happened in our daily life. Chapter 1 explores an advanced Gaussian Process (GP)-based optimization approach for the efficient geometry optimization of polypeptides, focusing on reducing computational costs associated with single-point energy (SPE) evaluations in traditional methods. By employing Gaussian Process Regression (GPR) as a surrogate model, the optimization steps are minimized through a surrogate potential energy surface (PES) generated from quantum mechanical data. The study assesses the performance of four kernel types—Matern, squared exponential, rational quadratic, and periodic—within multiple coordinate frameworks, including redundant and non-redundant internal coordinates and Coulombic coordinates. Results indicate that the periodic kernel combined with non-redundant delocalized internal coordinates is the most effective in reducing optimization steps, particularly suited to handle molecular structures with periodic characteristics. Additionally, the rational quadratic kernel shows promise when used with Coulombic coordinates, offering flexibility for functions with varying smoothness. Implemented in the mad-GP framework, this study provides insights into optimizing large biomolecules, such as polypeptides, with significant implications for computational chemistry and biomolecular modeling. We also compared the GP-optimized structures with the AlphaFold-predicted structures to assess their respective effectiveness in accurate structure prediction. This comparison provides insight into the reliability and applicability of each method for modeling polypeptide conformations. Chapter 2 investigates the interaction energies and energy decomposition of van der Waals (vdW) complexes between N6-methyladenosine (m6A) and tryptophan residues in YTH proteins, the readers of m6A modifications on mRNA. Given the role of m6A in cellular processes, structural insights into its interaction with YTH proteins could facilitate therapeutic advancements. We examined the effects of various chemical modifications on tryptophan residues (W465 and W470) in the YTH binding pocket, with the aim of enhancing the CH-π interactions with m6A through modified electron density. Using Density Functional Theory (DFT) and Symmetry-Adapted Perturbation Theory (SAPT), we explored the vdW interactions into electrostatic, dispersion, induction, and Pauli exchange components and identified London dispersion and electrostatics as dominant stabilizing forces. Correlations of these components with molecular descriptors such as polarizability and multipole moments further highlighted the effects of electronic properties on binding. Our results suggest that optimized tryptophan modifications could strengthen m6A recognition, potentially guiding the design of enhanced m6A-binding proteins for applications in RNA biology. Chapter 3 presents a computational analysis of reaction pathways and energy barriers on LiCoO₂ and TiO₂ surface models, exploring their role in promoting reactions critical to lithium-ion battery (LIB) performance and catalytic applications. For LiCoO₂, we examine the dissociation of H₂O₂. Using density functional theory (DFT) and climbing-image nudged elastic band (CI-NEB) calculations, we identified and characterized the elementary steps in the dissociation mechanism, and indicated that the reaction barriers are reduced in the presence of organic species. For TiO₂, we model the adsorption and dissociation of a Li(DME)₃ complex, exploring solvent dissociation and solvent exchange mechanisms in the context of DME ligands. Results show that the TiO₂ surface aids in stabilizing Li⁺ ions after solvent dissociation, and it favors a solvent-exchange pathway with a lower reaction barrier. These insights provide valuable mechanistic detail that help the design of materials. / Thesis (MS) — Boston College, 2024. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Chemistry.

2024 / Understanding the regulatory mechanisms of protein function is of major significance from both fundamental and biomedical points of view. Due to potential contributions from a multitude of physicochemical factors, analysis of regulatory mechanisms poses significant challenges to both experiments and computations. In this dissertation, a range of computational techniques have been developed and applied to better understand several representative regulatory mechanisms of protein function. One important regulatory mechanism of protein function emerged in recent studies concerns the hydration level of internal cavities. For example, water penetration was proposed to stabilize buried charges/dipoles, which play key roles in enzymes and bioenergetics systems. However, much uncertainty remains regarding the methodologies for describing the time scale and energetic driving forces for water penetration. Using extensive free energy simulations with polarizable force fields, we demonstrated that to properly describe the stability, hydration, dynamics, and therefore function of buried charges/dipoles, it is essential to explicitly include electronic polarization. Motivated by this observation, we have revised and implemented a grand canonical nonequilibrium candidate Monte Carlo approach to enable efficient sampling of cavity hydration level using a polarizable force field. These insights and method- ologies were essential to the analysis of the gating mechanism of the big potassium channel, in which the hydration level of the central hydrophobic cavity was proposed to regulate ion transport. Combined with nuclear magnetic resonance (NMR) spectroscopy, our enhanced sampling simulations also illustrated the roles and timescales of conformational change and internal hydration dynamics in determining the higher temperature-sensitivity of an engineered potassium channel. Another hallmark for biomolecules is that distal residues make significant cumulative contributions. However, their individual and specific roles remain difficult to predict and understand. We analyzed the contributions of second-shell residues in a metalloenzyme. By adopting a multifaceted approach that included both quantum mechanical and classical models, we probed the rate-limiting chemical step and structural properties of all relevant enzyme states. In combination with available experimental kinetics data, our results showed that mutations of those second-shell residues impact catalytic efficiency mainly by perturbation of the apo state and there- fore substrate binding, while they do not affect the ground state or transition state significantly. In another study, by examining a range of structural and dynamical properties in a transcription factor at both local and global scales in extensive molecular dynamics simulations, we showed that experimentally identified hotspot residues modulate allostery in distinct ways. The results motivated a thermodynamic model that qualitatively explained the broad distribution of hotspot residues observed in the experiment. We further demonstrated that the mutation effects of hotspot residues can be evaluated and ranked with functional free energy simulations. Collectively, these studies highlighted the power of integrating multiple computational approaches to better define the complex contributions of distal residues to function regulation.

Electrolytic solutions contain mobile ions that can pass current, and are essential components of any solution-phase electrochemical system. The Nernst–Planck–Poisson equations describe the electrodynamics and transport dynamics of electrolytic solutions. This thesis applies modern numerical and mathematical techniques in order to solve these equations, and hence determine the behaviour of electrochemical systems involving charge transport. The following systems are studied: a liquid junction where a concentration gradient causes charge transport; an ideally polarisable electrode where an applied potential difference causes charge transport; and an electrochemical cell where electrolysis causes charge transport. The nanometre Debye length and nanosecond Debye time scales are shown to control charge separation in electrolytic solutions. At equilibrium, charge separation is confined to within a Debye length scale of a charged electrode surface. Non-equilibrium charge separation is compensated in solution on a Debye time scale following a perturbation, whereafter electroneutrality dictates charge transport. The mechanism for the recovery of electroneutrality involves both migration and diffusion, and is non-linear for larger electrical potentials. Charge separation is an extremely important consideration on length scales comparable to the Debye length. The predicted features of capacitive charging and electrolysis at nanoelectrodes are shown to differ qualitatively from the behaviour of larger electrodes. Nanoscale charge separation can influence the behaviour of a larger system if it limits the overall rate of mass transport or electron transfer. This thesis advocates the use of numerical methods to solve the Nernst–Planck–Poisson equations, in order to avoid the simplifying approximations required by traditional analytical methods. As this thesis demonstrates, this methodology can reveal the behaviour of increasingly elaborate electrochemical systems, while illustrating the self-consistency and generality of fundamental theories concerning charge transport.

For the first time, successful synthesis and characterisation of the missing (Z)-isomer of thallium(I) methanediazotate has been accomplished, utilising low-temperature NMR monitoring analysis. The title compound was synthesised from N-methyl-N-nitrosourea and thallium(I) propoxide, under sub-ambient temperature conditions, as a highly moisture sensitive entity. Quantum chemical calculations, performed at the CCSD(T) level, depict excellent conformity to experimental results. Indeed, compared to its (E) counterpart, the formation of the title compound is thermodynamically less favoured, but preferred by means of kinetic control owing to a hindered isomerisation.

For the first time, successful synthesis and characterisation of the missing (Z)-isomer of thallium(I) methanediazotate has been accomplished, utilising low-temperature NMR monitoring analysis. The title compound was synthesised from N-methyl-N-nitrosourea and thallium(I) propoxide, under sub-ambient temperature conditions, as a highly moisture sensitive entity. Quantum chemical calculations, performed at the CCSD(T) level, depict excellent conformity to experimental results. Indeed, compared to its (E) counterpart, the formation of the title compound is thermodynamically less favoured, but preferred by means of kinetic control owing to a hindered isomerisation.

This thesis covers a range of methodologies to provide an account of the current (2010-2014) state of the art and to develop new methods for solubility prediction. We focus on predictions of intrinsic aqueous solubility, as this is a measure commonly used in many important industries including the pharmaceutical and agrochemical industries. These industries require fast and accurate methods, two objectives which are rarely complementary. We apply machine learning in chapters 4 and 5 suggesting methodologies to meet these objectives. In chapter 4 we look to combine machine learning, cheminformatics and chemical theory. Whilst in chapter 5 we look to predict related properties to solubility and apply them to a previously derived empirical equation. We also look at ab initio (from first principles) methods of solubility prediction. This is shown in chapter 3. In this chapter we present a proof of concept work that shows intrinsic aqueous solubility predictions, of sufficient accuracy to be used in industry, are now possible from theoretical chemistry using a small but diverse dataset. Chapter 6 provides a summary of our most recent research. We have begun to investigate predictions of sublimation thermodynamics. We apply quantum chemical, lattice minimisation and machine learning techniques in this chapter. In summary, this body of work concludes that currently, QSPR/QSAR methods remain the current state of the art for solubility prediction, although it is becoming possible for purely theoretical methods to achieve useful predictions of solubility. Theoretical chemistry can offer little useful additional input to informatics models for solubility predictions. However, theoretical chemistry will be crucial for enriching our understanding of the solvation process, and can have a beneficial impact when applied to informatics predictions of properties related to solubility.

The knowledge of the packing behaviour of small organic compounds in crystal lattices is of great importance for industries dealing with solid state materials. The properties of materials depend on how the molecules arrange themselves in a crystalline environment. Crystal structure prediction provides a theoretical approach through the application of computational strategies to seek possible crystal packing arrangements (or polymorphs) a compound may adopt. Based on the chemical diagrams, this thesis investigates polymorphism of several small organic compounds. Plausible crystal packings of those compounds are generated, and their lattice energies are minimised using molecular mechanics and/or quantum mechanics methods. Most of the work presented here is conducted using two software packages commercially available in this field, Polymorph Predictor of Materials Studio 4.0 and GRACE 1.0. In general, the computational techniques implemented in GRACE are very good at reproducing the geometries of the crystal structures corresponding to the experimental observations of the compounds, in addition to describing their solid state energetics correctly. Complementing the CSP results obtained using GRACE with isostructurality offers a route by which new potential polymorphs of the targeted compounds might be crystallised using the existing experimental data. Based on all calculations in this thesis, four new potential polymorphs for four different compounds, which have not yet been determined experimentally, are predicted to exist and may be obtained under the right crystallisation conditions. One polymorph is expected to crystallise under pressure. The remaining three polymorphs might be obtained by using a seeding technique or the utilisation of suitable tailor made additives.

A summary of research conducted since September 2007 at the University of Toronto in the laboratory of Professor Robert A. Batey is presented in this thesis, which is divided into four chapters. The first chapter contains a two-part introduction, where aryl- and aliphatic-Claisen rearrangements are discussed in part 1, and the nucleophilic addition of organoboron reagents to unsaturated C–N functionalities is described in part 2. Chapter 2 contains research involving synthetic and theoretical studies of aryl-Claisen rearrangements and other sigmatropic reactions. The work towards developing the lanthanide-catalyzed domino aryl-Claisen rearrangement for the synthesis of contiguous aryl–C(sp³) moieties is presented first. This is followed by computational studies involving E/Z-selectivity differences for the aryl-Claisen rearrangement, which was an issue noted for the domino aryl-Claisen reaction of a linear substrate. The mechanistic origins of E/Z-selectivity differences for the mono aryl-Claisen rearrangement, which was experimentally ambiguous for over 40 years, is resolved through computational methods. A theoretical analysis of selectivity differences for the allylic azide rearrangement is also described. The third section contains a discussion of Eu(fod)3-catalyzed aryl-Claisen rearrangements on vinyl bromide systems and preliminary studies involving application of the substrates in cross-coupling reactions, and other attempted mono- and domino sigmatropic rearrangements are presented in the fourth section. In chapter 3, the search for computational methods that can accurately predict experimental free energy of activation barriers for the aliphatic-Claisen rearrangement through benchmarking studies with a priori kinetic barrier and kinetic isotope effect data is described. Methods were found to predict new valid transition states and predict ΔG‡ values with a mean unsigned error of 0.3 kcal/mol relative to experimental values. In chapter 4, the development of new allylboration reaction is outlined, involving the double allylboration of nitriles and anhydrides, and initial studies towards the first aminoallylboration reactions of N-aluminoaldimines to form 1,2-diamines.

The rationalisation of drug potency using three-dimensional structures of protein-ligand complexes is a central paradigm in medicinal research. For over two decades, a major goal has been to find the rules that accurately relate the structure of any protein-ligand complex to its affinity. Addressing this problem is of great concern to the pharmaceutical industry, which uses virtual screens to computationally assay up to many millions of compounds against a protein target. A fast and trustworthy affinity estimator could potentially streamline the drug discovery process, reducing reliance on expensive wet lab experiments, speeding up the discovery of new hits and aiding lead optimization. Water plays a critical role in drug-protein interactions. To address the often ambiguous nature of water in binding sites, a water placement method was developed and found to be in good agreement with X-ray crystallography, neutron diffraction data and molecular dynamics simulations. The method is fast and has facilitated a large scale study of the statistics of water in ligand binding sites, as well as the creation of models pertaining to water binding free energies and displacement propensities, which are of particular interest to medicinal chemistry. Structure-based scoring functions employing the explicit water models were developed. Surprisingly, these attempts were no more accurate than the current state of the art, and the models suffered from the same inadequacies which have plagued all previous scoring functions. This suggests a unifying cause behind scoring function inaccuracy. Accordingly, mathematical analyses on the fundamental uncertainties in structure-based modelling were conducted. Using statistical learning theory and information theory, the existence of inherent errors in empirical scoring functions was proven. Among other results, it was found that even the very best generalised structure-based model is significantly limited in its accuracy, and protein-specific models are always likely to be better. The theoretical framework developed herein hints at modelling strategies that operate at the leading edge of achievable accuracy.

This thesis presents algorithm development in computational chemistry, and applies new computer science concepts to voltammetric simulation. To begin, this thesis discusses why algorithm development is necessary, and inherent problems found in commercial simulation solvers. As a result of this discussion, this thesis describes the need for simulators to keep abreast of recent computational developments. Algorithm development in this thesis is taken through stages. Chapter 3 applies known theory relating to the stripping voltammetry at a macroelectrode to the diffusional model of a microdisk, using finite difference and alternating direction implicit simulation techniques. Chapter 4 introduces the concept of parallel computing, and how computational hardware has developed recently to take advantage of out-of-order calculations, by processing them in parallel to reduce simulation time. The novel area of graphics card simulation for highly parallel algorithms is also explained in detail. Chapter 5 discusses the adaptation of voltammetric finite difference algorithms to a purely parallel format for simulation by explicit solution. Through explicit solution, finite difference algorithms are applied to electrode geometries which necessitate a three-dimensional solution – elliptical electrodes; square, rectangular, and microband electrodes; and dual microdisk electrodes in collector-generator mode. Chapter 6 introduces 'Random Walk' simulations, whereby individual particles in the simulation are modelled and their trajectories over time are calculated. The random walk technique in this thesis is improved for pure three-dimensional diffusion, and adapted to graphics cards, allowing up to a factor 4000 increase in speed over previous computational methods. This method is adapted to various systems of low concentration confined voltammetry (chapter 6.4) and single molecule detection, ultra low concentration cyclic voltammetry (chapter 6.5), and underpotential deposition of thallium on mobile silver nanoparticles (chapter 6.6). Overall, this thesis presents, and applies, a series of algorithm development concepts in computational electrochemistry.