Recent COMP News‎ > ‎

The OpenEye Outstanding Junior Faculty Award Winners for Denver (spring 2015)

posted Jan 28, 2015, 11:40 AM by Emilio Xavier Esposito
The COMP Division is excited to announce the OpenEye Outstanding Junior Faculty Award in Computational Chemistry winners for the Denver ACS meeting (spring 2015). Please visit the COMP award winners and the other excellent COMP posters at the COMP Poster Session on Tuesday, March 24, 2015 from 6pm to 8pm in Hall B2 of the Colorado Convention Center. More information about the OpenEye Outstanding Junior Faculty Award in Computational Chemistry can be found here.

Physically-motivated first-principles force fields for molecular simulation: Theory and applications
Jordan Schmidt, Jesse McDaniel. University of Wisconsin - Madison
Molecular simulations are widely used to predict and explain the structural, thermodynamic, and dynamic properties of condensed-phase chemical systems. As such, there has been considerable development of corresponding classical force fields. My group developed a novel approach for generating extremely accurate and transferable force fields exclusively on the basis of first-principles symmetry-adapted perturbation theory (SAPT) calculations. We use SAPT to calculate not only the total interaction energy, but also decompose this interaction into chemically-meaningful components: exchange repulsion, electrostatics, polarization, and dispersion. We exploit this energy decomposition to parameterize “physical-motivated” force fields which include, by construction, the right balance of all of the relevant “physics”, yielding robust and transferable force fields that yield the right answer, for the right reason. Since it does not rely on prior experimental data, we can use this approach even for systems that have not yet been previously characterized.

Development of electron-hole explicitly correlated wave function based method with pseudopotential theory for investigation of optical properties of quantum dot-protein complexes
Arindam ChakrabortySyracuse University
Firefly luciferase enzyme is an important candidate for bioluminescence resonance energy transfer in nano-bio systems. This work presents the theoretical investigation of the effect of protein corona formation by luciferase enzyme on the optical gap of CdSe quantum dot (QD). This is a computationally demanding calculation because of various factors including, the size of the QD (5nm Cd1159Se1183 cluster), number of luciferase molecules, and presence of explicit water molecules. We have combined the strengths of pseudopotential quantum mechanical calculations with molecular force fields to overcome the computational bottleneck. Specifically, the electronic structure of the QD was described by the CdSe pseudopotential and the resulting pseudopotential eigenvalue equation was solved using real-space localized basis functions. The effect of the adsorbed proteins on the optical gap was incorporated using electrostatic-embedded QM/MM approach, where the pseudopotential Hamiltonian was constructed in the presence of the electrostatic field generated by the partial atomic charges on the proteins. The electronic excitation in the QD was described using quasiparticle (electron-hole) representation and the eigenfunctions from the previous step were used for construction of the quasiparticle states. Electron-hole correlation was treated using the electron-hole explicitly correlated Hartree-Fock (eh-XCHF) method and quasiparticle states were used in the eh-XCHF method for calculation of the optical-gap of the QD-protein complex. Starting with the bare QD, the change in the optical gap due to addition of each enzyme molecule was computed. This process was continued until the formation of the protein corona was complete and additional enzyme binding was sterically unfavorable. The results provided the stoichiometry of the protein corona and quantified the effect of the number of bound protein molecules on the spectral-shift of the QD. The results from this work provide a route to correlate experimentally observed spectra with the structure of the protein corona.

Benchmarking the adsorption energies on carbon nanotubes
Daniel Smith, Konrad Patkowski. Auburn University
A thorough understanding of physisorption of small molecules on carbon nanotubes is of high importance to chemistry and material science. However, the adsorbate-nanotube interactions, dominated by London dispersion forces, pose a challenging problem for density functional theory (DFT). While many novel DFT variants that incorporate dispersion in some way have been proposed, none of them afford consistent accuracy for all systems. Thus, the selection of an optimal DFT variant has to be justified by a comparison to accurate interaction energies computed, for example, using coupled-cluster methods. In this work, we examined the adsorption of methane and carbon dioxide on graphene and carbon nanotubes. To this end, we first generated a set of 303 benchmark structures involving curved coronene as the nanotube model. For these systems, we computed accurate (errors below 0.1 kcal/mol at the minima) benchmark interaction energies using the composite MP2/CBS+ΔCCSD(T) approach. Subsequently, we evaluated the performance of a broad spectrum of modern DFT variants (including various flavors of the DFT+D approach, functionals optimized specifically for weak interactions such as M06-2X, the nonlocal VV10 functional, and the double-hybrid B2PLYP-D approach) on this benchmark set. For the methane-coronene complexes, several simple variants such as B3LYP-D3 afford good accuracy with mean relative errors in the 5-10% range. However, for the CO2-coronene system, no DFT variant tested exhibits similar accuracy, especially at distances shorter than the minimum. We show that the short-range deficiencies can be effectively removed by a refit of the damping parameters in the atom-pairwise dispersion expression, leading to a DFT-based approach that can be reliably used to examine the fragment size dependence of nanotube binding energy for larger finite models as well as infinite periodic nanotubes. On a more fundamental level, we investigate the distance dependence of the CCSD(T)-DFT difference that is to be recovered by the dispersion term and provide some constraints on the damping functions that are appropriate at all intermolecular separations.

RNA design rules through internet-scale social computing and high-throughput chemistry
Wipapat Kladwang, Daniel Cantu, Rhiju Das, Jeehyung Lee, Minjae Lee, Martin Azizyan, Limpaecher Alex, Adrien Treuille, Hanjoo Kim, Sungroh Yoon, EteRNA Participants. EteRNA project, Stanford, California; Carnegie Mellon University, Pittsburgh, Pennsylvania; Stanford University; Seoul National University, Seoul, South Korea
Self-assembling RNA molecules present compelling substrates for rational interrogation, control, and therapeutic intervention in living systems and disease. Computational methods for designing and predicting RNA secondary structures, switches, and 3D folds are playing fundamental roles in a new generation of RNA devices and therapeutics. However, imperfect in silico models – even at the secondary structure level – hinder the design of new RNAs that function properly when synthesized. This work presents a novel and potentially general approach to such empirical problems, the Massive Open Laboratory. The EteRNA project has connected 37,000 enthusiasts to RNA design ‘puzzles’ via an on-line interface. Uniquely, EteRNA participants not only manipulate simulated molecules, but also control a remote experimental pipeline for high-throughput RNA synthesis and structure mapping. Our initial results demonstrate how the EteRNA community leveraged dozens of cycles of continuous wet-lab feedback to learn strategies for solving in vitro RNA design problems on which automated methods fail. The top strategies – including several previously unrecognized ‘negative design rules’ – were distilled by machine learning into a new algorithm, EteRNABot. Over a rigorous one-year testing phase, both the EteRNA community and EteRNABot significantly outperformed prior algorithms in a dozen RNA secondary structure design tests, including the creation of dendrimer-like structures and scaffolds for small-molecule sensors. These results demonstrate for the first time that an online community can carry out large-scale experiments, hypothesis generation, and algorithm design to create practical advances in empirical science.