Thursday, November 18, 2021
Hendrik Heinz, Department of Chemical and Biological Engineering, University of Colorado Boulder
E-mail: [email protected]
Wonpil Im, Departments of Bioengineering and Computer Science, Lehigh University
E-mail: [email protected]
Ellad Tadmor, Department of Aerospace Engineering and Mechanics, University of Minnesota
E-mail: [email protected]
Title: Welcome and Introduction to Cyberloop Capabilities (Integration of IFF, CHARMM-GUI, OpenKIM) & Discussion
Abstract: We will introduce the capabilities of the individual platforms, including the current expansion of the Interface force field to new compounds (metals, oxides, gases, 2D materials), the new CHARMM-GUI Nanomaterials Modeler and its easy to use capabilities to build and run simulations of bionanomaterials, as well as advances in OpenKIM to integrate bonded force fields and develop validation standards suited for inorganic solids, soft matter interfaces, and force field comparisons.
Peter Coveney (Keynote), Department of Chemistry, University College London
E-mail: [email protected]
Title: Multiscale, Multiphysics Simulation Meets AI-Based Science for Advanced Materials Design
Abstract: The current approach to materials discovery and design remains dominated by experimental testing, often based on little more than trial and error. With the advent of ever more powerful computers, rapid, reliable, and reproducible computer simulations are beginning to represent a feasible alternative. As high performance computing reaches the exascale, exploiting the resources efficiently presents interesting challenges and opportunities. Multiscale modelling and simulation of materials are extremely promising candidates for exploiting these resources based on the assumption of a separation of scales in the architectures of nanomaterials. We apply multiscale techniques, from the electronic to the continuum scale, to the efficient design of a range of nanocomposites made from polymers and 2d-nanomaterials. We have also built an interoperable, open source and open development environment for validation, verification and uncertainty quantification which make such simulations actionable. A range of applications we are involved in cover not only advanced materials but fusion energy, weather/climate, biomedicine, epidemiology and human migration. Concurrently, we have developed hybrid machine learning (ML) and molecular dynamics (MD) workflows to accelerate drug discovery. The two in silico approaches each have their own complementary advantages and limitations. Here, we discuss an innovative cyberinfrastructure development that combines both approaches to accelerate drug discovery on supercomputers. Such workflows could also prove beneficial for advanced materials discovery.
Clare McCabe (Keynote), Department of Chemical & Biomolecular Engineering and Department of Chemistry, Vanderbilt University, Nashville TN 37235-1604
E-mail: [email protected]
Title: Understanding the Self-Assembly of Skin Lipids from Molecular Dynamics Simulations: A Multiscale Perspective
Abstract: The outermost layer of the skin (the stratum corneum) consists of skin cells embedded in a rich lipid matrix, whose primary role is to provide a barrier to foreign agents entering the body and to water leaving the body. This lipid system is unique in biological membranes in that it is composed of ceramides, cholesterol, and free fatty acids; phospholipids, which are the major components of most biological membranes, are completely absent. This unique composition enables the lipids of the stratum corneum to form highly organized lamella, which in turn are believed to control barrier function. Much is known about the nature of skin lipids from extensive experimental studies; nevertheless, a clear understanding of how and why these molecules assemble into the structures observed through microscopy and biophysical measurements does not yet exist. Molecular simulation can in principle provide molecular-level insight into membrane systems. However, the highly organized lamella observed in the stratum corneum make such studies challenging. Pre-assembled atomistic systems tend to be biased by the initial configuration because of low lipid mobility, and self-assembled structures require long timescales and large system sizes in order to study multilamellar structures, thus making self-assembly studies using atomistic models infeasible. Computationally efficient, coarse-grained models can be used to access these longer timescales and larger systems sizes required; however, coarse-grained models, by the nature of their construction, do not possess the same level of detail as atomistic systems, which may limit their ability to accurately model the structure and interactions in lipid membranes. As such, to reliably study the stratum corneum with simulation, an approach that combines the benefits of both atomistic and coarse-grained models is required. We will discuss our work in this area and our use of the MoSDeF framework to improve the efficiency and reproducibility of our simulations.
Lauren “Ren” Lopez, Northwestern University, Evanston, Illinois
E-mail: [email protected]
Title: Identification of Bioprivileged Molecules: Expansion of a Computational Approach to Broader Molecular Space
Abstract: As interest in biobased chemicals grows, and their application space expands, computational tools to navigate molecule space as a complement to experimental approaches are imperative. This work expands upon previous work that identified candidate bioprivileged molecules from the C6HxOy (C6) subspace. It refines the framework that was developed previously to better refine the molecules according to their biological origin and applies it to three new subspaces of chemical structure: C4HxOy (C4), C5HxOy (C5), and C7HxOy (C7). For C5 and C7, roughly the top 100 bioprivileged candidates were identified, and the enhanced framework was applied to recast slightly the previous list of the top 100 C6 molecules. In addition, all top candidates were analyzed for their key functional moieties using a random forest model, and this algorithm was applied to compare the functional group space occupied by bioprivileged molecules of various databases of molecules with a focus on evaluating how closely the molecules were aligned with those known to biology. Furthermore, with the present work's focus on automation and data science principles, the framework can be easily expanded to include other chemical formulae to screen for bioprivileged candidates. This in turn facilitates the retrosynthesis process inherent in the framework to identify those bioprivileged intermediates in other subspaces that lead to target molecules.
Jiali Gao (Keynote), Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, and
Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, China.
E-mail: [email protected]
Title: Direct Molecular Dynamics Using the Reactive Explicit Polarization (ReX-Pol) Method
Abstract: Traditionally, the study of reaction mechanisms of complex reaction systems has been performed on an individual basis. In this presentation, I will discuss a direct molecular dynamics (DMD) approach and the simulation program, CARNOT, in which all possible chemical reactions are simulated simultaneously at finite temperature and pressure conditions. A key concept of the fragment localized ab-initio molecular electronic structure method is to partition a large, chemically reactive system into molecular blocks on the fly in a dynamics simulation. The theory is called reactive explicit polarization (ReX-Pol) for reactive events. We fully explore the local nature of molecules, and the DMD approach can be applied to reactive systems consisting of an arbitrarily varying number of closed and open-shell species such as free radicals, zwitterions and separate ions found in combustion and other reactions. Employing the PW91 density functional theory and the 6-31+G(d) basis set, we will illustrate the capabilities of the CARNOT program by a combustion reaction consisting of more than 28,000 atoms. The trajectories reveal a range of mechanistic and dynamical events. Potential applications to biomolecular interactions and improvements will be discussed.
Donald G. Truhlar (Keynote), Department of Chemistry, University of Minnesota
E-mail: [email protected]
Title: Density Functional Theory and Multiconfiguration Pair-Density Functional Theory for Strongly Correlated Systems
Abstract: Most of the information we want to know about chemistry is in the electron density and electronic energy, and density functional theory has become the method of choice for electronic structure calculations on large and complex systems. A major challenge to density functional theory is the treatment of strongly correlated systems. Strongly correlated systems are those for which a single electron configuration does not provide a good zero-order treatment. Many transition metals species and transition states and most electronically excited states (even for the main group) are strongly correlated. I will discuss recent progress in Kohn-Sham density functional theory (KS-DFT) [1,2] and multiconfiguration pair-density functional theory (MC-PDFT) [3,4] that aims to meet this challenge. The accuracy of KS-DFT depends on the availability of accurate exchange-correlation functionals. By using the ingredients of kinetic energy density and range-separated nonlocal exchange, we have developed improved exchange-correlation functionals that are optimized to obtain a balanced treatment for weakly correlated and strongly correlated systems and that are optimized to obtain a balanced treatment of electronic excitation energies (simultaneously good performance for all types of excitation energies − valence, Rydberg, short-range charge transfer, and long-range charge transfer). Nevertheless, despite progress, KS-DFT remains better for weakly correlated systems than for strongly correlated ones. To obtain higher accuracy for transition metal chemistry and molecular excited states, we have developed MC-PDFT, which replaces the Slater determinant of KS-DFT with a multiconfiguration wave function and replaces the exchange-correlation functional of the up-spin and down-spin electron densities with an ontop functional of the total electron density and the on-top pair density. I will discuss the theory and applications of MC-PDFT. This work has been supported by research grants from the U.S. Department of Energy, the Air Force Office of Scientific Research, and the National Science Foundation. The work on MC-PDFT is a collaboration with co-principal-investigator Laura Gagliardi.
References:
[1] “Perspective: Kohn-Sham Density Functional Theory Descending a Staircase,” H. S. Yu, S. L. Li, and D. G. Truhlar, J. Chem. Phys. 145, 130901 (2016). doi.org/10.1063/1.4963168
[2] “Status and Challenges of Density Functional Theory,” P. Verma and D. G. Truhlar, Trends Chem. 2, 302-318 (2020). doi.org/10.1016/j.trechm.2020.02.005
[3] “Multiconfiguration Pair-Density Functional Theory: A New Way to Treat Strongly Correlated Systems,” L. Gagliardi, D. G. Truhlar, G. Li Manni, R. K. Carlson, C. E. Hoyer, and J. L. Bao, Acc. Chem. Res. 50, 66-73 (2017). doi.org/10.1021/acs.accounts.6b00471
[4] “Multiconfiguration Pair-Density Functional Theory,” P. Sharma, J. J. Bao, D. G. Truhlar, and L. Gagliardi, Annu. Rev. Phys. Chem. 72, 541-564 (2021). doi.org/10.1146/annurev-physchem-090419-043839
Friday, November 19, 2021
Yeol Kyo Choi, Department of Biological Sciences, Lehigh University
E-mail: [email protected]
Title: CHARMM-GUI Nanomaterial Modeler for Modeling and Simulation of Nanomaterial Systems & Convergence Discussion
Abstract: Molecular modeling and simulation are invaluable tools for nanoscience that predict mechanical, physicochemical, and thermodynamic properties of nanomaterials and provide molecular-level insight into underlying mechanisms. However, building nanomaterial-containing systems remains challenging due to the lack of reliable and integrated cyberinfrastructures. Here, we present Nanomaterial Modeler in CHARMM-GUI, a web-based cyberinfrastructure that provides an automated process to generate various nanomaterial models, associated topology, and configuration files to perform state-of-the-art molecular dynamics simulations using most simulation packages. The transferability of nanomaterial models among the simulation programs and effect of various Lennard-Jones cut-off methods on structural and interfacial properties of nanomaterials are discussed. Furthermore, Nanomaterial Modeler's capabilities will be illustrated by molecular dynamics simulation of selected representative model systems.
Moon-Ki Choi, Department of Aerospace Engineering and Mechanics, University of Minnesota
E-mail: [email protected]
Title: Automated Comparison of Bonded Force Fields and Reactive Interatomic Potentials in OpenKIM
Abstract: Bonded force fields (FFs), such as IFF [1] and CHARMM [2,3], are widely used in bio- and nano-materials simulations. To date such FFs have not been directly compared with the reactive interatomic potentials (IPs) commonly used in computational materials science, such as EAM, Tersoff, REBO, ReaxFF, and so on. To enable such comparisons, the OpenKIM framework [4,5], which archives and tests reactive IPs, is being extended to support bonded FFs. As first examples, property calculations (aka "KIM Tests") for graphite and molybdenum disulfide [6] are being developed with input from the CHARM-GUI Nanomaterial Modeler [7], including (1) equilibrium lattice cohesive energy at 0 K; (2) cleavage energy at 0 K; (3) lattice constant and cohesive energy at room temperature and pressure; (4) cleavage energy at room temperate and pressure. The Tests at finite temperature and pressure employ the `kim-convergence` Python package, developed by the KIM team for automated run-time control by determining equilibration and steady-state convergence to a desired confidence level. This is necessary to enable KIM Tests to run with arbitrary FFs and IPs without manual intervention. Computed properties are integrated within the OpenKIM repository (https://openkim.org) where they can be viewed and compared.
[1] H. Heinz, T.-J. Lin, R. K. Mishra, F. S. Emami. Langmuir 29 (2013) 1754-1765
[2] J. Huang, A. D. MacKerell. J. Comput. Chem. 34(25) (2013) 2135-2145
[3] J. Sunhwan, T. Kim, V. G. Iyer, W. Im. et al. J. Comput. Chem. 29(11) (2008) 1859-1865
[4] E. B. Tadmor, R. S. Elliott, J. P. Sethna, R. E. Miller, C. A. Becker. JOM, 63 (2011) 17.
[5] D. S. Karls, M. Bierbaum, A. A. Alemi, R. S. Elliott, J. P. Sethna. E. B. Tadmor. J. Chem. Phys. 153 (2020) 064104.
[6] J. Liu, J. Zeng, C. Zhu, J. Miao, W. Heinz. Chem. Sci. 11 (2020)
8708-8722.
[7] Y. K. Choi, N. R. Kern, S. Kim et al. J. Chem. Theo. Comp., (2021)
submitted.
Krishan Kanhaiya, Department of Chemical and Biological Engineering, University of Colorado Boulder
E-mail: [email protected]
Title: Simulation of Metals, Oxides, Hydroxides, and Biointerfaces up to the Large Nanometer Scale & IFF Convergence Discussion
Abstract: The simulation of metals, oxides, and hydroxides is important to guide in the design of therapeutics, catalysts, cement, ceramics, biocomposites and glasses. Here we explain the Interface force field (IFF) for ~30 metals, oxides, and hydroxides (all fcc metals, Fe, Cr, NiO, CaO, MgO, β-Ni(OH)2, β-Ca(OH)2, α-Al2O3, α-Cr2O3, α-Fe2O3, and SiO2) using non-bonded only potentials that involve Lennard-Jones and Coulomb terms. The LJ parameters for metals outperform existing parameters and density functionals in terms of surface and interfacial properties at high speed. Atomic charges for oxides and hydroxides are close to internal dipole moments even when explicit covalent bonds are excluded. The non-bonded models for oxides and hydroxides are inherently reactive, enabling changes in morphology and defects. The models achieve quantitative agreement with experiment in lattice parameters (<1% deviation), surface energies to the extent known, and in bulk modulus (<10% deviation). All force field parameters and models can be used with existing parameters for inorganic compounds, solvents, polymers and biomolecules via compatibility with CHARMM, CVFF, AMBER, OPLS-AA, PCFF, and COMPASS. The metal, oxide, and hydroxide parameters are available with 12-6 and 9-6 Lennard-Jones potential options and integrated in CHARMM-GUI Nanomaterial Modeler. The models can be used for simulations at length scales from atoms to 1 µm^3 and outperform popular atomistic models (Pedone et. al. and ReaxFF) in speed, reliability, and compatibility. The incorporation of fully bonded parameters for oxides and hydroxides is also possible.
Sapna Sarupria (Keynote), Department of Chemistry, University of Minnesota
E-mail: [email protected]
Title: Leveraging Advanced Simulations and Machine Learning to Overcome Free Energy Barriers in Molecular Systems
Abstract: Many interesting and important processes in molecular systems such as protein folding, the transition of liquid-to-solid and chemical reactions involve overcoming free energy barriers. Often the free energy barriers are large, and this makes it almost impossible to sample these processes through straightforward molecular dynamics simulations. Consequently, several methods have been developed to address this limitation. In our research, we use a family of methods called path sampling techniques. We have integrated data science methods and path sampling techniques to enable large-scale simulations of the molecular processes of interest. In addition, we are using machine learning approaches to better understand the data generated from these calculations. In my talk, I will discuss existing challenges in studying these important yet hard-to-sample molecular processes and hope to initiate conversations of the potential role of machine learning and integration of cyberinfrastructure in solving some of these challenges.
Massimiliano “Max” Bonomi (Keynote), Structural Bioinformatics Unit, Department of Structural Biology and Chemistry; CNRS UMR 3528; C3BI, CNRS USR 3756; Institut Pasteur, Paris, France
E-mail: [email protected]
Title: Integrative Computational/Experimental Approaches to Study the Interaction of Peptides with Mineral Surfaces
Abstract: Elucidation of the structure and interactions of proteins at native mineral interfaces is key to understanding how biological systems regulate the formation of hard tissue structures. In addition, understanding how these same proteins interact with non-native mineral surfaces has important implications for the design of medical and dental implants, chromatographic supports, diagnostic tools, and a host of other applications. In this presentation, I will illustrate how SNa15, a peptide derived from the hydroxyapatite (HAP) recognition domain of the biomineralization protein statherin, interacts with HAP, silica (SiO2), and titania (TiO2) mineral surfaces. Our results [1] show that SNa15 adopts an α-helical conformation when adsorbed to HAP and TiO2, but the helix largely unravels upon adsorption to SiO2. Interactions with HAP are mediated in general by acidic and some basic amino acids, although the specific amino acids involved in direct surface interaction vary with surface. This study has been carried out using an integrative computational and experimental approach [2] that combines solid-state NMR spectroscopy with enhanced-sampling molecular dynamics simulations to obtain high-resolution insights into adsorption of proteins on interfaces. This integrative approach is implemented in the open-source, freely-available PLUMED library (www.plumed.org) [3] and can be readily used to determine structural and dynamical properties of conformationally heterogeneous systems by integrating different types of experimental data into molecular simulations.
[1] E. L. Buckle, A. Prakash, M. Bonomi, J. Sampath, J. Pfaendtner, G. P. Drobny. J. Am. Chem. Soc. 141 (2019) 1998
[2] M. Bonomi, C. Camilloni, A. Cavalli, M. Vendruscolo. Sci. Adv. 2 (2016) e1501177
[3] Bonomi, et al., Nat. Methods 16 (2019) 670
Dongyue Liang, Boston University and University of Chicago, Institute of Molecular Engineering
E-mail: [email protected]
Title: Force Field Development and Molecular Dynamics Studies at LiCoO2-Water Interface
Abstract: Nanomaterials are playing increasingly important roles in modern industry, and their applications are expanding rapidly. This, however, also leads to their exposure to the biological environment, and their environmental impacts are not well-understood. A class of oxide nanomaterials, LiNixMnyCo1-x-yO2 ("NMC"), which has great potential for energy storage applications, is found to release ions in solution and negatively impact the growth and survival of bacteria. To enable the interfacial adsorption and kinetic studies, we applied quantum mechanics and classical modeling methods to develop a force field suitable for characterizing LiCoO2-water interface. Further molecular dynamics studies at the interface highlighted the contribution of water molecules to the interfacial adsorption behaviors of small molecules.
Jim Pfaendtner (Keynote), Department of Chemical Engineering, University of Washington, Seattle, WA 98195
E-mail: [email protected]
Title: Opportunities and Challenges in Simulating Biomolecules at Interfaces
Abstract: Molecular simulation of interfacial phenomena of biomolecules presents an enormous opportunity to advance our understanding of a host of important systems spanning human health, advanced materials and energy. MD simulations promise nanoscale resolution in length and time along with detailed mechanistic insights. However, many obstacles remain in expanding this research subfield to create more user friendly and mainstream tools. This talk will highlight recent examples of biomolecule/surface simulations from my group, primarily focused around biomineralizing peptides, and conclude with a discussion of priority research directions for expanding access and reliability of these simulations in the future.
____________________________________________________________________________________________________________________________________________
Hendrik Heinz, Department of Chemical and Biological Engineering, University of Colorado Boulder
E-mail: [email protected]
Wonpil Im, Departments of Bioengineering and Computer Science, Lehigh University
E-mail: [email protected]
Ellad Tadmor, Department of Aerospace Engineering and Mechanics, University of Minnesota
E-mail: [email protected]
Title: Welcome and Introduction to Cyberloop Capabilities (Integration of IFF, CHARMM-GUI, OpenKIM) & Discussion
Abstract: We will introduce the capabilities of the individual platforms, including the current expansion of the Interface force field to new compounds (metals, oxides, gases, 2D materials), the new CHARMM-GUI Nanomaterials Modeler and its easy to use capabilities to build and run simulations of bionanomaterials, as well as advances in OpenKIM to integrate bonded force fields and develop validation standards suited for inorganic solids, soft matter interfaces, and force field comparisons.
Peter Coveney (Keynote), Department of Chemistry, University College London
E-mail: [email protected]
Title: Multiscale, Multiphysics Simulation Meets AI-Based Science for Advanced Materials Design
Abstract: The current approach to materials discovery and design remains dominated by experimental testing, often based on little more than trial and error. With the advent of ever more powerful computers, rapid, reliable, and reproducible computer simulations are beginning to represent a feasible alternative. As high performance computing reaches the exascale, exploiting the resources efficiently presents interesting challenges and opportunities. Multiscale modelling and simulation of materials are extremely promising candidates for exploiting these resources based on the assumption of a separation of scales in the architectures of nanomaterials. We apply multiscale techniques, from the electronic to the continuum scale, to the efficient design of a range of nanocomposites made from polymers and 2d-nanomaterials. We have also built an interoperable, open source and open development environment for validation, verification and uncertainty quantification which make such simulations actionable. A range of applications we are involved in cover not only advanced materials but fusion energy, weather/climate, biomedicine, epidemiology and human migration. Concurrently, we have developed hybrid machine learning (ML) and molecular dynamics (MD) workflows to accelerate drug discovery. The two in silico approaches each have their own complementary advantages and limitations. Here, we discuss an innovative cyberinfrastructure development that combines both approaches to accelerate drug discovery on supercomputers. Such workflows could also prove beneficial for advanced materials discovery.
Clare McCabe (Keynote), Department of Chemical & Biomolecular Engineering and Department of Chemistry, Vanderbilt University, Nashville TN 37235-1604
E-mail: [email protected]
Title: Understanding the Self-Assembly of Skin Lipids from Molecular Dynamics Simulations: A Multiscale Perspective
Abstract: The outermost layer of the skin (the stratum corneum) consists of skin cells embedded in a rich lipid matrix, whose primary role is to provide a barrier to foreign agents entering the body and to water leaving the body. This lipid system is unique in biological membranes in that it is composed of ceramides, cholesterol, and free fatty acids; phospholipids, which are the major components of most biological membranes, are completely absent. This unique composition enables the lipids of the stratum corneum to form highly organized lamella, which in turn are believed to control barrier function. Much is known about the nature of skin lipids from extensive experimental studies; nevertheless, a clear understanding of how and why these molecules assemble into the structures observed through microscopy and biophysical measurements does not yet exist. Molecular simulation can in principle provide molecular-level insight into membrane systems. However, the highly organized lamella observed in the stratum corneum make such studies challenging. Pre-assembled atomistic systems tend to be biased by the initial configuration because of low lipid mobility, and self-assembled structures require long timescales and large system sizes in order to study multilamellar structures, thus making self-assembly studies using atomistic models infeasible. Computationally efficient, coarse-grained models can be used to access these longer timescales and larger systems sizes required; however, coarse-grained models, by the nature of their construction, do not possess the same level of detail as atomistic systems, which may limit their ability to accurately model the structure and interactions in lipid membranes. As such, to reliably study the stratum corneum with simulation, an approach that combines the benefits of both atomistic and coarse-grained models is required. We will discuss our work in this area and our use of the MoSDeF framework to improve the efficiency and reproducibility of our simulations.
Lauren “Ren” Lopez, Northwestern University, Evanston, Illinois
E-mail: [email protected]
Title: Identification of Bioprivileged Molecules: Expansion of a Computational Approach to Broader Molecular Space
Abstract: As interest in biobased chemicals grows, and their application space expands, computational tools to navigate molecule space as a complement to experimental approaches are imperative. This work expands upon previous work that identified candidate bioprivileged molecules from the C6HxOy (C6) subspace. It refines the framework that was developed previously to better refine the molecules according to their biological origin and applies it to three new subspaces of chemical structure: C4HxOy (C4), C5HxOy (C5), and C7HxOy (C7). For C5 and C7, roughly the top 100 bioprivileged candidates were identified, and the enhanced framework was applied to recast slightly the previous list of the top 100 C6 molecules. In addition, all top candidates were analyzed for their key functional moieties using a random forest model, and this algorithm was applied to compare the functional group space occupied by bioprivileged molecules of various databases of molecules with a focus on evaluating how closely the molecules were aligned with those known to biology. Furthermore, with the present work's focus on automation and data science principles, the framework can be easily expanded to include other chemical formulae to screen for bioprivileged candidates. This in turn facilitates the retrosynthesis process inherent in the framework to identify those bioprivileged intermediates in other subspaces that lead to target molecules.
Jiali Gao (Keynote), Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, and
Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, China.
E-mail: [email protected]
Title: Direct Molecular Dynamics Using the Reactive Explicit Polarization (ReX-Pol) Method
Abstract: Traditionally, the study of reaction mechanisms of complex reaction systems has been performed on an individual basis. In this presentation, I will discuss a direct molecular dynamics (DMD) approach and the simulation program, CARNOT, in which all possible chemical reactions are simulated simultaneously at finite temperature and pressure conditions. A key concept of the fragment localized ab-initio molecular electronic structure method is to partition a large, chemically reactive system into molecular blocks on the fly in a dynamics simulation. The theory is called reactive explicit polarization (ReX-Pol) for reactive events. We fully explore the local nature of molecules, and the DMD approach can be applied to reactive systems consisting of an arbitrarily varying number of closed and open-shell species such as free radicals, zwitterions and separate ions found in combustion and other reactions. Employing the PW91 density functional theory and the 6-31+G(d) basis set, we will illustrate the capabilities of the CARNOT program by a combustion reaction consisting of more than 28,000 atoms. The trajectories reveal a range of mechanistic and dynamical events. Potential applications to biomolecular interactions and improvements will be discussed.
Donald G. Truhlar (Keynote), Department of Chemistry, University of Minnesota
E-mail: [email protected]
Title: Density Functional Theory and Multiconfiguration Pair-Density Functional Theory for Strongly Correlated Systems
Abstract: Most of the information we want to know about chemistry is in the electron density and electronic energy, and density functional theory has become the method of choice for electronic structure calculations on large and complex systems. A major challenge to density functional theory is the treatment of strongly correlated systems. Strongly correlated systems are those for which a single electron configuration does not provide a good zero-order treatment. Many transition metals species and transition states and most electronically excited states (even for the main group) are strongly correlated. I will discuss recent progress in Kohn-Sham density functional theory (KS-DFT) [1,2] and multiconfiguration pair-density functional theory (MC-PDFT) [3,4] that aims to meet this challenge. The accuracy of KS-DFT depends on the availability of accurate exchange-correlation functionals. By using the ingredients of kinetic energy density and range-separated nonlocal exchange, we have developed improved exchange-correlation functionals that are optimized to obtain a balanced treatment for weakly correlated and strongly correlated systems and that are optimized to obtain a balanced treatment of electronic excitation energies (simultaneously good performance for all types of excitation energies − valence, Rydberg, short-range charge transfer, and long-range charge transfer). Nevertheless, despite progress, KS-DFT remains better for weakly correlated systems than for strongly correlated ones. To obtain higher accuracy for transition metal chemistry and molecular excited states, we have developed MC-PDFT, which replaces the Slater determinant of KS-DFT with a multiconfiguration wave function and replaces the exchange-correlation functional of the up-spin and down-spin electron densities with an ontop functional of the total electron density and the on-top pair density. I will discuss the theory and applications of MC-PDFT. This work has been supported by research grants from the U.S. Department of Energy, the Air Force Office of Scientific Research, and the National Science Foundation. The work on MC-PDFT is a collaboration with co-principal-investigator Laura Gagliardi.
References:
[1] “Perspective: Kohn-Sham Density Functional Theory Descending a Staircase,” H. S. Yu, S. L. Li, and D. G. Truhlar, J. Chem. Phys. 145, 130901 (2016). doi.org/10.1063/1.4963168
[2] “Status and Challenges of Density Functional Theory,” P. Verma and D. G. Truhlar, Trends Chem. 2, 302-318 (2020). doi.org/10.1016/j.trechm.2020.02.005
[3] “Multiconfiguration Pair-Density Functional Theory: A New Way to Treat Strongly Correlated Systems,” L. Gagliardi, D. G. Truhlar, G. Li Manni, R. K. Carlson, C. E. Hoyer, and J. L. Bao, Acc. Chem. Res. 50, 66-73 (2017). doi.org/10.1021/acs.accounts.6b00471
[4] “Multiconfiguration Pair-Density Functional Theory,” P. Sharma, J. J. Bao, D. G. Truhlar, and L. Gagliardi, Annu. Rev. Phys. Chem. 72, 541-564 (2021). doi.org/10.1146/annurev-physchem-090419-043839
Friday, November 19, 2021
Yeol Kyo Choi, Department of Biological Sciences, Lehigh University
E-mail: [email protected]
Title: CHARMM-GUI Nanomaterial Modeler for Modeling and Simulation of Nanomaterial Systems & Convergence Discussion
Abstract: Molecular modeling and simulation are invaluable tools for nanoscience that predict mechanical, physicochemical, and thermodynamic properties of nanomaterials and provide molecular-level insight into underlying mechanisms. However, building nanomaterial-containing systems remains challenging due to the lack of reliable and integrated cyberinfrastructures. Here, we present Nanomaterial Modeler in CHARMM-GUI, a web-based cyberinfrastructure that provides an automated process to generate various nanomaterial models, associated topology, and configuration files to perform state-of-the-art molecular dynamics simulations using most simulation packages. The transferability of nanomaterial models among the simulation programs and effect of various Lennard-Jones cut-off methods on structural and interfacial properties of nanomaterials are discussed. Furthermore, Nanomaterial Modeler's capabilities will be illustrated by molecular dynamics simulation of selected representative model systems.
Moon-Ki Choi, Department of Aerospace Engineering and Mechanics, University of Minnesota
E-mail: [email protected]
Title: Automated Comparison of Bonded Force Fields and Reactive Interatomic Potentials in OpenKIM
Abstract: Bonded force fields (FFs), such as IFF [1] and CHARMM [2,3], are widely used in bio- and nano-materials simulations. To date such FFs have not been directly compared with the reactive interatomic potentials (IPs) commonly used in computational materials science, such as EAM, Tersoff, REBO, ReaxFF, and so on. To enable such comparisons, the OpenKIM framework [4,5], which archives and tests reactive IPs, is being extended to support bonded FFs. As first examples, property calculations (aka "KIM Tests") for graphite and molybdenum disulfide [6] are being developed with input from the CHARM-GUI Nanomaterial Modeler [7], including (1) equilibrium lattice cohesive energy at 0 K; (2) cleavage energy at 0 K; (3) lattice constant and cohesive energy at room temperature and pressure; (4) cleavage energy at room temperate and pressure. The Tests at finite temperature and pressure employ the `kim-convergence` Python package, developed by the KIM team for automated run-time control by determining equilibration and steady-state convergence to a desired confidence level. This is necessary to enable KIM Tests to run with arbitrary FFs and IPs without manual intervention. Computed properties are integrated within the OpenKIM repository (https://openkim.org) where they can be viewed and compared.
[1] H. Heinz, T.-J. Lin, R. K. Mishra, F. S. Emami. Langmuir 29 (2013) 1754-1765
[2] J. Huang, A. D. MacKerell. J. Comput. Chem. 34(25) (2013) 2135-2145
[3] J. Sunhwan, T. Kim, V. G. Iyer, W. Im. et al. J. Comput. Chem. 29(11) (2008) 1859-1865
[4] E. B. Tadmor, R. S. Elliott, J. P. Sethna, R. E. Miller, C. A. Becker. JOM, 63 (2011) 17.
[5] D. S. Karls, M. Bierbaum, A. A. Alemi, R. S. Elliott, J. P. Sethna. E. B. Tadmor. J. Chem. Phys. 153 (2020) 064104.
[6] J. Liu, J. Zeng, C. Zhu, J. Miao, W. Heinz. Chem. Sci. 11 (2020)
8708-8722.
[7] Y. K. Choi, N. R. Kern, S. Kim et al. J. Chem. Theo. Comp., (2021)
submitted.
Krishan Kanhaiya, Department of Chemical and Biological Engineering, University of Colorado Boulder
E-mail: [email protected]
Title: Simulation of Metals, Oxides, Hydroxides, and Biointerfaces up to the Large Nanometer Scale & IFF Convergence Discussion
Abstract: The simulation of metals, oxides, and hydroxides is important to guide in the design of therapeutics, catalysts, cement, ceramics, biocomposites and glasses. Here we explain the Interface force field (IFF) for ~30 metals, oxides, and hydroxides (all fcc metals, Fe, Cr, NiO, CaO, MgO, β-Ni(OH)2, β-Ca(OH)2, α-Al2O3, α-Cr2O3, α-Fe2O3, and SiO2) using non-bonded only potentials that involve Lennard-Jones and Coulomb terms. The LJ parameters for metals outperform existing parameters and density functionals in terms of surface and interfacial properties at high speed. Atomic charges for oxides and hydroxides are close to internal dipole moments even when explicit covalent bonds are excluded. The non-bonded models for oxides and hydroxides are inherently reactive, enabling changes in morphology and defects. The models achieve quantitative agreement with experiment in lattice parameters (<1% deviation), surface energies to the extent known, and in bulk modulus (<10% deviation). All force field parameters and models can be used with existing parameters for inorganic compounds, solvents, polymers and biomolecules via compatibility with CHARMM, CVFF, AMBER, OPLS-AA, PCFF, and COMPASS. The metal, oxide, and hydroxide parameters are available with 12-6 and 9-6 Lennard-Jones potential options and integrated in CHARMM-GUI Nanomaterial Modeler. The models can be used for simulations at length scales from atoms to 1 µm^3 and outperform popular atomistic models (Pedone et. al. and ReaxFF) in speed, reliability, and compatibility. The incorporation of fully bonded parameters for oxides and hydroxides is also possible.
Sapna Sarupria (Keynote), Department of Chemistry, University of Minnesota
E-mail: [email protected]
Title: Leveraging Advanced Simulations and Machine Learning to Overcome Free Energy Barriers in Molecular Systems
Abstract: Many interesting and important processes in molecular systems such as protein folding, the transition of liquid-to-solid and chemical reactions involve overcoming free energy barriers. Often the free energy barriers are large, and this makes it almost impossible to sample these processes through straightforward molecular dynamics simulations. Consequently, several methods have been developed to address this limitation. In our research, we use a family of methods called path sampling techniques. We have integrated data science methods and path sampling techniques to enable large-scale simulations of the molecular processes of interest. In addition, we are using machine learning approaches to better understand the data generated from these calculations. In my talk, I will discuss existing challenges in studying these important yet hard-to-sample molecular processes and hope to initiate conversations of the potential role of machine learning and integration of cyberinfrastructure in solving some of these challenges.
Massimiliano “Max” Bonomi (Keynote), Structural Bioinformatics Unit, Department of Structural Biology and Chemistry; CNRS UMR 3528; C3BI, CNRS USR 3756; Institut Pasteur, Paris, France
E-mail: [email protected]
Title: Integrative Computational/Experimental Approaches to Study the Interaction of Peptides with Mineral Surfaces
Abstract: Elucidation of the structure and interactions of proteins at native mineral interfaces is key to understanding how biological systems regulate the formation of hard tissue structures. In addition, understanding how these same proteins interact with non-native mineral surfaces has important implications for the design of medical and dental implants, chromatographic supports, diagnostic tools, and a host of other applications. In this presentation, I will illustrate how SNa15, a peptide derived from the hydroxyapatite (HAP) recognition domain of the biomineralization protein statherin, interacts with HAP, silica (SiO2), and titania (TiO2) mineral surfaces. Our results [1] show that SNa15 adopts an α-helical conformation when adsorbed to HAP and TiO2, but the helix largely unravels upon adsorption to SiO2. Interactions with HAP are mediated in general by acidic and some basic amino acids, although the specific amino acids involved in direct surface interaction vary with surface. This study has been carried out using an integrative computational and experimental approach [2] that combines solid-state NMR spectroscopy with enhanced-sampling molecular dynamics simulations to obtain high-resolution insights into adsorption of proteins on interfaces. This integrative approach is implemented in the open-source, freely-available PLUMED library (www.plumed.org) [3] and can be readily used to determine structural and dynamical properties of conformationally heterogeneous systems by integrating different types of experimental data into molecular simulations.
[1] E. L. Buckle, A. Prakash, M. Bonomi, J. Sampath, J. Pfaendtner, G. P. Drobny. J. Am. Chem. Soc. 141 (2019) 1998
[2] M. Bonomi, C. Camilloni, A. Cavalli, M. Vendruscolo. Sci. Adv. 2 (2016) e1501177
[3] Bonomi, et al., Nat. Methods 16 (2019) 670
Dongyue Liang, Boston University and University of Chicago, Institute of Molecular Engineering
E-mail: [email protected]
Title: Force Field Development and Molecular Dynamics Studies at LiCoO2-Water Interface
Abstract: Nanomaterials are playing increasingly important roles in modern industry, and their applications are expanding rapidly. This, however, also leads to their exposure to the biological environment, and their environmental impacts are not well-understood. A class of oxide nanomaterials, LiNixMnyCo1-x-yO2 ("NMC"), which has great potential for energy storage applications, is found to release ions in solution and negatively impact the growth and survival of bacteria. To enable the interfacial adsorption and kinetic studies, we applied quantum mechanics and classical modeling methods to develop a force field suitable for characterizing LiCoO2-water interface. Further molecular dynamics studies at the interface highlighted the contribution of water molecules to the interfacial adsorption behaviors of small molecules.
Jim Pfaendtner (Keynote), Department of Chemical Engineering, University of Washington, Seattle, WA 98195
E-mail: [email protected]
Title: Opportunities and Challenges in Simulating Biomolecules at Interfaces
Abstract: Molecular simulation of interfacial phenomena of biomolecules presents an enormous opportunity to advance our understanding of a host of important systems spanning human health, advanced materials and energy. MD simulations promise nanoscale resolution in length and time along with detailed mechanistic insights. However, many obstacles remain in expanding this research subfield to create more user friendly and mainstream tools. This talk will highlight recent examples of biomolecule/surface simulations from my group, primarily focused around biomineralizing peptides, and conclude with a discussion of priority research directions for expanding access and reliability of these simulations in the future.
____________________________________________________________________________________________________________________________________________