Deep learning for presumed probability density function models
Not Available
SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Not Available
Explore the source record for details and available documents.
Plane wave pseudopotential (PWP) density functional theory (DFT) calculation is the most widely used method for material simulations, but its absolute speed stagnated due to the inability to use large scale CPU based computers. By a drastic redesign of the algorithm, and moving all the major computation parts into GPU, we have reached a speed of 12 s per molecular dynamics (MD) step for a 512 atom system using 256 GPU cards. This is about 20 times faster than the CPU version of the code regardless of the number of CPU cores used. Our tests and analysis on different GPU platforms and configurations shed lights on the optimal GPU deployments for PWP-DFT calculations. An 1800 step MD simulation is used to study the liquid phase properties of GaInP.
Modeling large datasets through Multivariate Functional Approximations (MFA) provide an elegant way to handle many visualization and scientific analysis workflows. The process necessitates scalable data partitioning methods to compute MFA representations efficiently without compromising the accuracy or continuity of the reconstructed solution. We propose a domain -decomposed method for computing the MFA with B -spline bases, which reduces the total work per task and uses a restricted Additive Schwarz (RAS) method to converge the control point data degrees -of -freedom along subdomain boundaries. We provide an in-depth analysis of the parallel approach with domain decomposition solvers, aiming to minimize local subdomain error residuals and recover high -order continuity at subdomain interfaces with appropriate choices of knot overlaps. The communication cost, determined by the overlap regions in the RAS implementation, is optimized to recover the numerical error profile of the single subdomain case. Our proposed method stands in contrast to previous methods, which typically only recover either C 0 or at best C 1 continuity for arbitrary B -spline degree expansions, or those that require post -processing to blend discontinuities in the reconstructed data. We demonstrate the effectiveness of our approach using analytical and real -world datasets in 1D, 2D, and 3D through both strong and weak scaling studies. The performance results indicate that the overall cost of computing the approximation is directly proportional to the underlying nearest -neighbor communication implementation, and is only weakly dependent on the overlap region size that determines the size of the messages. This finding underscores the efficiency and scalability of our proposed method, making it a promising solution for handling large datasets in scientific workflows.
Explore the source record for details and available documents.
Many pathogenic bacteria have sophisticated mechanisms to interfere with the mammalian immune response. These include the disruption of host extracellular ATP levels that, in humans, is tightly regulated by the nucleoside triphosphate diphosphohydrolase family (NTPDases). NTPDases are found almost exclusively in eukaryotes, the notable exception being their presence in some pathogenic prokaryotes. To address the function of bacterial NTPDases, we describe the structures of an NTPDase from the pathogen Legionella pneumophila (Lpg1905/Lp1NTPDase) in its apo state and in complex with the ATP analog AMPPNP and the subtype-specific NTPDase inhibitor ARL 67156. Lp1NTPDase is structurally and catalytically related to eukaryotic NTPDases and the structure provides a basis for NTPDase-specific inhibition. Furthermore, we demonstrate that the activity of Lp1NTPDase correlates directly with intracellular replication of Legionella within macrophages. Collectively, these findings provide insight into the mechanism of this enzyme and highlight its role in host-pathogen interactions.
The B cell antigen receptor (BCR) plays an essential role in all phases of B cell development. Here we show that the extracellular domains of murine and human Ig{beta} form an I-set immunoglobulin-like structure with an interchain disulfide between cysteines on their G strands. Structural and sequence analysis suggests that Ig{alpha} displays a similar fold as Ig{beta}. An Ig{alpha}{beta} heterodimer model was generated based on the unique disulfide-bonded Ig{beta} dimer. Solution binding studies showed that the extracellular domains of Ig{alpha}{beta} preferentially recognize the constant region of BCR with chain specificity, suggesting a role for Ig{alpha}{beta} to enhance BCR chain signaling. Cluster mutations on Ig{alpha}, Ig{beta}, and a membrane-bound form of immunoglobulin (mIgM) based on the structural model identified distinct areas of potential contacts involving charged residues on both subunits of the coreceptor and the C4 domain of mIgM. These studies provide the first structural model for understanding BCR function.
Budding of HIV-1 requires the binding of the PTAP late domain of the Gag p6 protein to the UEV domain of the TSG101 subunit of ESCRT-I. The normal function of this motif in cells is in receptor downregulation. Here, we report the 1.41.6 {angstrom} structures of the human TSG101 UEV domain alone and with wild-type and mutant HIV-1 PTAP and Hrs PSAP nonapeptides. The hydroxyl of the Thr or Ser residue in the P(S/T)AP motif hydrogen bonds with the main chain of Asn69. Mutation of the Asn to Pro, blocking the main-chain amide, abrogates PTAP motif binding in vitro and blocks budding of HIV-1 from cells. N69P and other PTAP binding-deficient alleles of TSG101 did not rescue HIV-1 budding. However, the mutant alleles did rescue downregulation of endogenous EGF receptor. This demonstrates that the PSAP motif is not rate determining in EGF receptor downregulation under normal conditions.
A comprehensive ab initio study of periodic actinide–ligand bonding trends for trivalent actinides is performed. Relativistic density functional theory (DFT) and complete active-space (CAS) self-consistent field wavefunction calculations are used to dissect the chemical bonding in the [AnCl 6 ] 3– , [An(CN) 6 ] 3– , [An(NCS) 6 ] 3– , [An(S 2 PMe 2 ) 3 ], [An(DPA) 3 ] 3– , and [An(HOPO)] – series of actinide (An = U–Es) complexes. Except for some differences for the early actinide complexes with DPA, bond orders and excess 5f-shell populations from donation bonding show qualitatively similar trends in 5f n active-space CAS vs DFT calculations. The influence of spin–orbit coupling on donation bonding is small for the tested systems. Along the actinide series, chemically soft vs chemically harder ligands exhibit clear differences in bonding trends. There are pronounced changes in the 5f populations when moving from Pu to Am or Cm, which correlate with previously noted “breaks” in chemical trends. As a result, bonding involving 5f becomes very weak beyond Cm/Bk. We propose that Cm(III) is a borderline case among the trivalent actinides that can be meaningfully considered to be involved in ground-state 5f covalent bonding.
We compare several established approaches for propagating wave functions and calculating hopping probabilities within the fewest switches surface hopping (FSSH) algorithm for difficult cases with many electronic states and many trivial crossings. If only a single time step (Δt c ) is employed, we find that no published approach can accurately capture the dynamics correctly unless Δt c → 0 (which is not computationally feasible). If multiple time steps are employed, for a fixed classical time step (Δt c ), a robust scheme can be found for dynamically choosing quantum time steps (δt q1 and δt q2 ) and calculating hopping probabilities so that one can systematically reduce all errors and achieve maximally efficient accuracy; scattering calculations confirm that one can choose a fairly large classical time step. Furthermore, the robust scheme presented here uses both the “local diabatic” and adiabatic interpolation and thus borrows elements from both the Granucci/Persico and Meek/Levine algorithms. Our findings should be broadly applicable in the future.
A concise synthetic method has been developed to access functionalized naphtho[2,3]porphyrins through combining two sequence reactions involving a Heck–electrocyclization–aromatization sequence and a Wittig–Knovenegal sequence. Using this method, mononaphtho[2,3]porphyrin (NP-1), opp-dinaphtho[2,3]porphyrin (NP-2), and push–pull naphtho[2,3]porphyrin (NP-3) have been prepared. These naphtho[2,3]porphyrins displayed interesting optical and electrochemical properties. As a result, excellent efficiencies of singlet oxygen generation were obtained for these naphtho[2,3]porphyrins.
Two uncharacterized enzymes from the amidohydrolase superfamily belonging to cog1228 were cloned, expressed, and purified to homogeneity. The two proteins, Sgx9260c (gi|44242006) and Sgx9260b (gi|44479596), were derived from environmental DNA samples originating from the Sargasso Sea. The catalytic function and substrate profiles for Sgx9260c and Sgx9260b were determined using a comprehensive library of dipeptides and N-acyl derivative of L-amino acids. Sgx9260c catalyzes the hydrolysis of Gly-L-Pro, L-Ala-L-Pro, and N-acyl derivatives of L-Pro. The best substrate identified to date is N-acetyl-L-Pro with a value of k{sub cat}/K{sub m} of 3 x 10{sup 5} M{sup -1} s{sup -1}. Sgx9260b catalyzes the hydrolysis of L-hydrophobic L-Pro dipeptides and N-acyl derivatives of L-Pro. The best substrate identified to date is N-propionyl-L-Pro with a value of k{sub cat}/K{sub m} of 1 x 10{sup 5} M{sup -1} s{sup -1}. Three-dimensional structures of both proteins were determined by X-ray diffraction methods (PDB codes 3MKV and 3FEQ). These proteins fold as distorted ({beta}/{alpha})8-barrels with two divalent cations in the active site. The structure of Sgx9260c was also determined as a complex with the N-methylphosphonate derivative of L-Pro (PDB code 3N2C). In this structure the phosphonate moiety bridges the binuclear metal center, and one oxygen atom interacts with His-140. The {alpha}-carboxylate of the inhibitor interacts with Tyr-231. The proline side chain occupies a small substrate binding cavity formed by residues contributed from the loop that follows {beta}-strand 7 within the ({beta}/{alpha})8-barrel. A total of 38 other proteins from cog1228 are predicted to have the same substrate profile based on conservation of the substrate binding residues. The structure of an evolutionarily related protein, Cc2672 from Caulobacter crecentus, was determined as a complex with the N-methylphosphonate derivative of L-arginine (PDB code 3MTW).
N/A
We describe the use of Cl K-edge X-ray Absorption Spectroscopy (XAS) and both ground state and time-dependent hybrid density functional theory (DFT) to probe electronic structure and determine the degree of orbital mixing in M-Cl bonds for (C5Me5)2MCl2 (M = Ti, 1; Zr, 2; Hf, 3; Th, 4; and U, 5), where we can directly compare a class of structurally similar compounds for d- and f-elements. We report direct experimental evidence for covalency in M-Cl bonding, including actinides, and offer insight into the relative roles of the valence f- and dorbitals in these systems. The Cl K-edge XAS data for the group IV transition metals, 1 – 3, show slight decreases in covalency in M-Cl bonding with increasing principal quantum number, in the order Ti > Zr > Hf. The percent Cl 3p character per M-Cl bond was experimentally determined to be 25, 23, and 22% per M-Cl bond for 1-3, respectively. For actinides, we find a shoulder on the white line for (C5Me5)2ThCl2, 4, and distinct, but weak pre-edge features for 2 (C5Me5)2UCl2, 5. The percent Cl 3p character in Th-Cl bonds in 4 was determined to be 14 %, with high uncertainty, while the U-Cl bonds in 5 contains 9 % Cl 3p character. The magnitudes of both values are approximately half what was observed for the transition metal complexes in this class of bent metallocene dichlorides. Using the hybrid DFT calculations as a guide to interpret the experimental Cl K-edge XAS, these experiments suggest that when evaluating An- Cl bonding, both 5f- and 6d-orbitals should be considered. For (C5Me5)2ThCl2, the calculations and XAS indicate that the 5f- and 6d-orbitals are nearly degenerate and heavily mixed. In contrast, the 5f- and 6d-orbitals in (C5Me5)2UCl2 are no longer degenerate, and fall in two distinct energy groupings. The 5f-orbitals are lowest in energy and split into a 5-over-2 pattern with the high lying U 6d-orbitals split in a 4-over-1 pattern, the latter of which is similar to the dorbital splitting in group IV transition metal (C5R5)2MCl2 (R = H, Me) compounds. Time dependent-DFT (TD-DFT) was used to calculate the energies and intensities of Cl 1s transitions into empty metal based orbitals containing Cl 3p character, and provide simulated Cl K-edge XAS spectra for 1 - 4. However, for 5, which has two unpaired electrons, analogous information was obtained from transition dipole calculations using ground state Kohn-Sham orbitals. The simulations provide additional confidence in the interpretation of spectra based on ground state calculations. Overall, this study demonstrates that Cl K-edge XAS and DFT calculations represent powerful tools that can be used to evaluate electronic structure and covalency in actinide metal-ligand bonding. In addition, these results provide a framework that can be used in future studies to evaluate actinide covalency in compounds that contain transuranic elements.
Abstract not provided
We have performed a detailed analysis of the magnetic (collinear and non-collinear) order and the atomic and electron structures of UO2, PuO2 and UN on the basis of density functional theory with the Hubbard electron correlation correction (DFT + U). We have shown that the 3-k magnetic structure of UO2 is the lowest in energy for the Hubbard parameter value of U = 4.6 eV (and J = 0.5 eV) consistent with experiments when Dudarev’s formalism is used. In contrast to UO2, UN and PuO2 show no trend for a distortion towards rhombohedral structure and, thus, no complex 3-k magnetic structure is to be anticipated in these materials.