Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Pattern Recognition”

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.

At least 1,513 records · Page 84

Structural determinants of nuclear export signal orientation in binding to exportin CRM1

The Chromosome Region of Maintenance 1 (CRM1) protein mediates nuclear export of hundreds of proteins through recognition of their nuclear export signals (NESs), which are highly variable in sequence and structure. The plasticity of the CRM1-NES interaction is not well understood, as there are many NES sequences that seem incompatible with structures of the NES-bound CRM1 groove. Crystal structures of CRM1 bound to two different NESs with unusual sequences showed the NES peptides binding the CRM1 groove in the opposite orientation (minus) to that of previously studied NESs (plus). A comparison of minus and plus NESs identified structural and sequence determinants for NES orientation. The binding of NESs to CRM1 in both orientations results in a large expansion in NES consensus patterns and therefore a corresponding expansion of potential NESs in the proteome.

59 BASIC BIOLOGICAL SCIENCES↗

Investigation about relationships between the symmetries of ferroelectric crystal Ca{sub 0.28}Ba{sub 0.72}Nb{sub 2}O{sub 6} and second-harmonic patterns

The broadband quasi-phase matching (QPM) process in a uniaxial ferroelectric crystal Ca{sub 0.28}Ba{sub 0.72}Nb{sub 2}O{sub 6} (CBN-28) was demonstrated with the second-harmonic wavelength range from 450 nm to 650 nm, and the relationship between the symmetries of CBN-28 and the second-harmonic patterns was experimentally and theoretically investigated based on the random anti-parallel domains in the crystal and QPM conditions. The dependences of frequency-doubled patterns on the wavelength and anisotropy of the nonlinear crystal were also studied, and the frequency-doubled photons were found to be trapped on circles. By analyzing the light-matter interacting Hamiltonians, the trapping force for second-harmonic photons was found to be centripetal and tunable by the fundamental lasers, and the variation tendencies of the rotational velocity of second-harmonic generation photons could also be predicated. The results indicate that the CBN-28 ferroelectric crystal is a promising nonlinear optical material for the generation of broadband frequency-doubled waves, and the analysis on centripetal force based on the interaction Hamiltonians may provide a novel recognition for the investigation of QPM process to be further studied.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Dynamical Signatures of Thermotoga maritima Maltose-Binding Proteins Affected by Ligand Binding

Functional segregation among protein isoforms depends on the interplay of their overall structures and the molecular dynamics of these structures. Thermotoga maritima maltose-binding protein (tmMBP) isoforms show size-dependent differential binding of maltose and malto-oligomers while maintaining remarkable fold conservation. This differential behavior needs detailed characterization in native-like aqueous conditions to understand the effects of protein dynamics on ligand binding and recognition. Small-angle neutron scattering (SANS), neutron spin echo (NSE) spectroscopy, and dynamic light scattering (DLS) were used in conjunction with previously published computational molecular dynamics (MD) simulations to understand the dynamic behavior of tmMBPs experimentally. SANS provided information on the overall structure of the molecules, while NSE was used to determine the dynamics in the nanosecond time scale. Both tmMBP2 and tmMBP3 have a bidomain architecture linked with a flexible hinge, with the binding pocket sitting in the cleft between the two domains. tmMBP2 and tmMBP3 showed different solution dynamics, with the translational and rotational components dominating the dynamics of both systems, resulting in a clear differentiation of their diffusion pattern. A faster dynamics component was also observed and was attributed to segmental dynamics. Differences observed between the ligand-free (apo) and ligand-bound (holo) states of the two proteins are attributed to conformational entropy. Our results highlight the intricacies of how structure and dynamics can together shape binding to a repertoire of substrates in structurally similar proteins.

Diffusion↗

Effect of glycosylation on an immunodominant region in the V1V2 variable domain of the HIV-1 envelope gp120 protein

Heavy glycosylation of the envelope (Env) surface subunit, gp120, is a key adaptation of HIV-1; however, the precise effects of glycosylation on the folding, conformation and dynamics of this protein are poorly understood. Here we explore the patterns of HIV-1 Env gp120 glycosylation, and particularly the enrichment in glycosylation sites proximal to the disulfide linkages at the base of the surface-exposed variable domains. To dissect the influence of glycans on the conformation these regions, we focused on an antigenic peptide fragment from a disulfide bridge-bounded region spanning the V1 and V2 hyper-variable domains of HIV-1 gp120. We used replica exchange molecular dynamics (MD) simulations to investigate how glycosylation influences its conformation and stability. Simulations were performed with and without N-linked glycosylation at two sites that are highly conserved across HIV-1 isolates (N156 and N160); both are contacts for recognition by V1V2-targeted broadly neutralizing antibodies against HIV-1. Glycosylation stabilized the pre-existing conformations of this peptide construct, reduced its propensity to adopt other secondary structures, and provided resistance against thermal unfolding. Simulations performed in the context of the Env trimer also indicated that glycosylation reduces flexibility of the V1V2 region, and provided insight into glycan-glycan interactions in this region. These stabilizing effects were influenced by a combination of factors, including the presence of a disulfide bond between the Cysteines at 131 and 157, which increased the formation of beta-strands. Together, these results provide a mechanism for conservation of disulfide linkage proximal glycosylation adjacent to the variable domains of gp120 and begin to explain how this could be exploited to enhance the immunogenicity of those regions. Furthermore, these studies suggest that glycopeptide immunogens can be designed to stabilize the most relevant Env conformations to focus the immune response on key neutralizing epitopes.

59 BASIC BIOLOGICAL SCIENCES↗

Recognition of a Key Anchor Residue by a Conserved Hydrophobic Pocket Ensures Subunit Interface Integrity in DNA Clamps

Sliding clamp proteins encircle duplex DNA and are involved in processive DNA replication and the DNA damage response. Clamp proteins are ring-shaped oligomers (dimers or trimers) and are loaded onto DNA by an ATP-dependent clamp loader complex that ruptures the interface between two adjacent subunits. Here we measured the solution dynamics of the human clamp protein, proliferating cell nuclear antigen, by monitoring the change in the fluorescence of a site-specifically labeled. To unravel the origins of clamp subunit interface stability, we carried out comprehensive comparative analysis of the interfaces of seven sliding clamps. Here, we used computational modeling (molecular dynamic simulations and MM/GBSA binding energy decomposition analyses) to identify conserved networks of hydrophobic residues critical for clamp stability and ring-opening dynamics. The hydrophobic network is shared among clamp proteins and exhibits a “key in a keyhole” pattern where a bulky aromatic residue from one clamp subunit is anchored into a hydrophobic pocket of the opposing subunit. Bioinformatics and dynamic network analyses showed that this oligomeric latch is conserved across DNA sliding clamps from all domains of life and dictates the dynamics of clamp opening and closing.

59 BASIC BIOLOGICAL SCIENCES↗

Structure-Based Analysis of Toxoplasma gondii Profilin: A Parasite-Specific Motif Is Required for Recognition by Toll-Like Receptor 11

Profilins promote actin polymerization by exchanging ADP for ATP on monomeric actin and delivering ATP-actin to growing filament barbed ends. Apicomplexan protozoa such as Toxoplasma gondii invade host cells using an actin-dependent gliding motility. Toll-like receptor (TLR) 11 generates an innate immune response upon sensing T. gondii profilin (TgPRF). The crystal structure of TgPRF reveals a parasite-specific surface motif consisting of an acidic loop, followed by a long {beta}-hairpin. A series of structure-based profilin mutants show that TLR11 recognition of the acidic loop is responsible for most of the interleukin (IL)-12 secretion response to TgPRF in peritoneal macrophages. Deletion of both the acidic loop and the {beta}-hairpin completely abrogates IL-12 secretion. Insertion of the T. gondii acidic loop and {beta}-hairpin into yeast profilin is sufficient to generate TLR11-dependent signaling. Substitution of the acidic loop in TgPRF with the homologous loop from the apicomplexan parasite Cryptosporidium parvum does not affect TLR11-dependent IL-12 secretion, while substitution with the acidic loop from Plasmodium falciparum results in reduced but significant IL-12 secretion. We conclude that the parasite-specific motif in TgPRF is the key molecular pattern recognized by TLR11. Unlike other profilins, TgPRF slows nucleotide exchange on monomeric rabbit actin and binds rabbit actin weakly. The putative TgPRF actin-binding surface includes the {beta}-hairpin and diverges widely from the actin-binding surfaces of vertebrate profilins.

59 BASIC BIOLOGICAL SCIENCES↗

Structure and Functional Characterization of the RNA-Binding Element of the NLRX1 Innate Immune Modulator

Mitochondrial NLRX1 is a member of the family of nucleotide-binding domain and leucine-rich-repeat-containing proteins (NLRs) that mediate host innate immunity as intracellular surveillance sensors against common molecular patterns of invading pathogens. NLRX1 functions in antiviral immunity, but the molecular mechanism of its ligand-induced activation is largely unknown. The crystal structure of the C-terminal fragment (residues 629975) of human NLRX1 (cNLRX1) at 2.65 {angstrom} resolution reveals that cNLRX1 consists of an N-terminal helical (LRRNT) domain, central leucine-rich repeat modules (LRRM), and a C-terminal three-helix bundle (LRRCT). cNLRX1 assembles into a compact hexameric architecture that is stabilized by intersubunit and interdomain interactions of LRRNT and LRRCT in the trimer and dimer components of the hexamer, respectively. Furthermore, we find that cNLRX1 interacts directly with RNA and supports a role for NLRX1 in recognition of intracellular viral RNA in antiviral immunity.

59 BASIC BIOLOGICAL SCIENCES↗

Study on small-strain behaviours of methane hydrate sandy sediments using discrete element method

Methane hydrate bearing soil has attracted increasing interest as a potential energy resource where methane gas can be extracted from dissociating hydrate-bearing sediments. Seismic testing techniques have been applied extensively and in various ways, to detect the presence of hydrates, due to the fact that hydrates increase the stiffness of hydrate-bearing sediments. With the recognition of the limitations of laboratory and field tests, wave propagation modelling using Discrete Element Method (DEM) was conducted in this study in order to provide some particle-scale insights on the hydrate-bearing sandy sediment models with pore-filling and cementation hydrate distributions. The relationship between shear wave velocity and hydrate saturation was established by both DEM simulations and analytical solutions. Obvious differences were observed in the dependence of wave velocity on hydrate saturation for these two cases. From the shear wave velocity measurement and particle-scale analysis, it was found that the small-strain mechanical properties of hydrate-bearing sandy sediments are governed by both the hydrate distribution patterns and hydrate saturation.

58 GEOSCIENCES↗

Oxygen, Magnesium, and Aluminum Isotopes in the Ivuna CAI: Re-Examining High-Temperature Fractionations in CI Chondrites

CI chondrites are thought to approximate the bulk solar system composition since they closely match the composition of the solar photosphere. Thus, chemical differences between a planetary object and the CI composition are interpreted to result from fractionations of a CI starting composition. This interpretation is often made despite the secondary mineralogy of CI chondrites, which resulted from low-T aqueous alteration on the parent asteroid(s). Prevalent alteration and the relatively large uncertainties in the photospheric abundances (approx. +/-5-10%) permit chemical fractionation of CI chondrites from the bulk solar system, if primary chondrules and/or CAIs have been altered beyond recognition. Isolated olivine and pyroxene grains that range from approx. 5 microns to several hundred microns have been reported in CI chondrites, and acid residues of Orgueil were found to contain refractory oxides with oxygen isotopic compositions matching CAIs. However, the only CAI found to be unambiguously preserved in a CI chondrite was identified in Ivuna. The Ivuna CAI's primary mineralogy, small size (approx.170 microns), and fine-grained igneous texture classify it as a compact type A. Aqueous alteration infiltrated large portions of the CAI, but other regions remain pristine. The major primary phases are melilite (Ak 14-36 ), grossmanite (up to 20.8 wt.% TiO 2 ), and spinel. Both melilite and grossmanite have igneous textures and zoning patterns. An accretionary rim consists primarily of olivine (Fa 2-17 ) and low-Ca pyroxene (Fs 2-10 ), which could be either surviving CI2 material or a third lithology.

Frank, D. R.↗

Fleetite, Cu 2 RhIrSb 2 , a New Species of Platinum-Group Mineral from the Miass Placer Zone, Southern Urals, Russia

Fleetite, Cu 2 RhIrSb 2 , a new species of platinum-group mineral (PGM), was discovered intergrown with an Os–Ir–Ru alloy in the Miass Placer Zone (Au–PGE), southern Urals, Russia. A single grain 50 μm across was found. Osmium, ruthenium, and iridium are the main associated minerals; also present are Pt–Fe alloys, laurite, Sb-rich irarsite, Rh-rich tolovkite, kashinite, anduoite, ferronickelplatinum, heazlewoodite, PGE-bearing pentlandite and digenite, as well as micrometric inclusions of forsterite (Fo 93.7 ), chromite–magnesiochromite, and Mg-rich edenite. In reflected light, fleetite is light gray; it is opaque, isotropic, non-pleochroic, and non-bireflectant. We report reflectance values measured in air. A mean of seven point-analyses (wavelength-dispersive spectrometry) gave Cu 13.93, Ni 8.60, Fe 0.10, Ir 28.07, Rh 7.91, Ru 1.96, Sb 39.28, total 99.85 wt.%, corresponding to (Cu 1.41 Ni 0.58 Fe 0.01 ) Σ2.00 (Rh 0.49 Ni 0.36 Ru 0.12 ) Σ0.97 Ir 0.95 Sb 2.08 on the basis of six atoms per formula unit, taking into account the structural results. The calculated density is 10.83 g/cm 3 . Single-crystal X-ray studies show that fleetite is cubic, space group Fdm (#227), a = 11.6682(8) Å, V = 1588.59(19) Å 3 , and Z = 16. A least-squares refinement of X-ray powder-diffraction data gave a = 11.6575(5) Å and V = 1584.22(19) Å 3 . The strongest five reflections in the powder pattern [d in Å(I)(hkl)] are: 6.70(75)(111), 4.13(100)(220), 3.52(30)(311), 2.380(50)(422), 2.064(40)(440). Results of synchrotron micro-Laue diffraction experiments are consistent [a = 11.66(2) Å]. The crystal structure of fleetite was solved and refined to R = 0.0340 based upon 153 reflections with F o > 4σ(F o ). It is isotypic with Pd 11 Bi 2 Se 2 and best described as intermetallic, with all metal atoms in 12-fold coordination. Fleetite and other late exotic phases were formed by reaction of the associated alloy phases with a fluid phase enriched in Sb, As, and S in circulation in the cooling ophiolite source-rock. The mineral is named after Michael E. Fleet (1938–2017) in recognition of his significant contributions to the Earth Sciences.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Multiscale simulations reveal TDP-43 molecular-level interactions driving condensation

The RNA-binding protein TDP-43 is associated with mRNA processing and transport from the nucleus to the cytoplasm. TDP-43 localizes in the nucleus as well as accumulating in cytoplasmic condensates such as stress granules. Aggregation and formation of amyloid-like fibrils of cytoplasmic TDP-43 are hallmarks of numerous neurodegenerative diseases, most strikingly present in >90% of amyotrophic lateral sclerosis (ALS) patients. If excessive accumulation of cytoplasmic TDP-43 causes, or is caused by, neurodegeneration is presently not known. In this work, we use molecular dynamics simulations at multiple resolutions to explore TDP-43 self- and cross-interaction dynamics. A full-length molecular model of TDP-43, all 414 amino acids, was constructed from select structures of the protein functional domains (N-terminal domain, and two RNA recognition motifs, RRM1 and RRM2) and modeling of disordered connecting loops and the low complexity glycine-rich C-terminus domain. All-atom CHARMM36m simulations of single TDP-43 proteins served as guides to construct a coarse-grained Martini 3 model of TDP-43. The Martini model and a coarser implicit solvent Cα model, optimized for disordered proteins, were subsequently used to probe TDP-43 interactions; self-interactions from single-chain full-length TDP-43 simulations, cross-interactions from simulations with two proteins and simulations with assemblies of dozens to hundreds of proteins. Our findings illustrate the utility of different modeling scales for accessing TDP-43 molecular-level interactions and suggest that TDP-43 has numerous interaction preferences or patterns, exhibiting an overall strong, but dynamic, association and driving the formation of biomolecular condensates.

59 BASIC BIOLOGICAL SCIENCES↗

Insights into the dual cleavage activity of the GH16 laminarinase enzyme class on β-1,3 and β-1,4 glycosidic bonds

Glycoside hydrolases (GH) are involved in the degradation of a wide diversity of carbohydrates and present several biotechnological applications. Many GH families are composed of enzymes with a single well-defined specificity. In contrast, enzymes from the GH16 family can act on a range of different polysaccharides, including β-glucans and galactans. SCLam, a GH16 member derived from a soil metagenome, an endo-β-1,3(4)-glucanase (EC 3.2.1.6), can cleave both β-1,3 and β-1,4 glycosidic bonds in glucans, such as laminarin, barley β-glucan, and cello-oligosaccharides. A similar cleavage pattern was previously reported for other GH16 family members. However, the molecular mechanisms for this dual cleavage activity on (1,3)- and (1,4)-β-D-glycosidic bonds by laminarinases have not been elucidated. In this sense, we determined the X-ray structure of a presumably inactive form of SCLam co-crystallized with different oligosaccharides. The solved structures revealed general bound products that are formed due to residual activities of hydrolysis and transglycosylation. Biochemical and biophysical analyses and molecular dynamics simulations help to rationalize differences in activity towards different substrates. Our results depicted a bulky aromatic residue near the catalytic site critical to select the preferable configuration of glycosidic bonds in the binding cleft. Altogether, these data contribute to understanding the structural basis of recognition and hydrolysis of β-1,3 and β-1,4 glycosidic linkages of the laminarinase enzyme class, which is valuable for future studies on the GH16 family members and applications related to biomass conversion into feedstocks and bioproducts.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Real Time Intelligent Target Detection and Analysis with Machine Vision

We present an algorithm for detecting a specified set of targets for an Automatic Target Recognition (ATR) application. ATR involves processing images for detecting, classifying, and tracking targets embedded in a background scene. We address the problem of discriminating between targets and nontarget objects in a scene by evaluating 40x40 image blocks belonging to an image. Each image block is first projected onto a set of templates specifically designed to separate images of targets embedded in a typical background scene from those background images without targets. These filters are found using directed principal component analysis which maximally separates the two groups. The projected images are then clustered into one of n classes based on a minimum distance to a set of n cluster prototypes. These cluster prototypes have previously been identified using a modified clustering algorithm based on prior sensed data. Each projected image pattern is then fed into the associated cluster's trained neural network for classification. A detailed description of our algorithm will be given in this paper. We outline our methodology for designing the templates, describe our modified clustering algorithm, and provide details on the neural network classifiers. Evaluation of the overall algorithm demonstrates that our detection rates approach 96% with a false positive rate of less than 0.03%.

Howard, Ayanna↗

Diagnostic Clast-Texture Criteria for Recognition of Impact Deposits

It is difficult to find definitive evidence for impact in the geological record because there are many endogenous geological processes that can produce diamictites similar to those generated by impact ejecta. The classic impact criteria of shock fabrics in certain minerals, and iridium layers, for example, may be either difficult to find, or long-since erased from the impact site (shock fabrics also anneal with time). It is important to be able to recognize impact-generated materials in order to understand earth's crustal development and biological evolution. In future exploration of Mars and other solar-system bodies, recognition of impact materials will be important for elucidating planetary evolution, planetary volatile inventories, and exobiological issues. The cobble depicted is typical of many that have been found in diamictite deposits in Belize generated by the Chicxulub K-T impact event. The pebbles are roughly-hewn in general shape with smoothed corners and edges. Surfaces are almost uniformly frosted (on both protuberances and hollows), but some asperities are glazed. Optical microscopy and thin-section petrographic microscopy reveal the frosting to be only a few microns thick, with a well-defined granular structure; grains are the same size as those composing the bulk of the limestone, but their clearer appearance may represent annealing. One or two adjacent pebble faces are often decorated with striated gouges and closely-spaced hemispherical depressions representing indentation hollows produced by well-rounded impacting clasts of up to 0.5 cm in diameter. Some of the impactors are still embedded in the cobble surface. Non-destructive x-ray diffraction techniques showed the impactors to be of the same mineralogy as the target cobble. We believe this unusual glazing and frosting to be related to the impact event, but this must be reconciled with its survival for over 60 my. since it is composed of one of the most alterable substances, CaCO3. We focus, however, on the non-fractured rounded pebbles that appear to have impacted the larger cobble as a swarm of clasts, probably encountered during ballistic flight. It cannot be defined whether the cobble swept through the pebble cloud, or the pebbles rained upon a slowermoving cobble. Two interesting questions arise: (1) Where did large numbers of such well rounded pebbles come from in the ejecta curtain?, and (2) How did they embed themselves in a nominally brittle rock without suffering damage? Are the well-rounded pebbles crystalline (devitrified) melt spherules? Further investigations are in progress. If the cobble was cold and brittle, impact of well-rounded pebbles would have produced Hertzian fracture patterns (a) in the virtually elastically isotropic cobble target. For penetration depths of about 0.5 of the impactor diameter, the pebbles would require a relative impact velocity sufficient to cause pebble fragmentation and crushing, and the development of a complex Hertzian-Boussinesq fracture field (b) involving deep fracturing and lateral surface spalling (impact velocities > 50 m/s). The existing relationship could only evolve by the impact of cold, hard pebbles into a soft, plastic cobble surface (c). For limestone to have been plastic, it must have been at elevated temperature, but to prevent calcification of the material, the ambient pressure must also have been elevated. This would be possible either in the impact's gas plume, or within the confines of a thin aerodynamically-produced shock bow generated by supersonic ballistic motion of the cobble. In the latter case, it is implied that the cobble swept through the pebbles rather than vice-versa. Although there are high T-P conditions associated with volcanism, such textures have not been reported on volcaniclastic materials, nor from other high T-P environments; e.g., metamorphic (as far as the authors are aware). We propose that this very easily recognizable embedding and indentation surface texture can be used as a diagnostic criterion for the recognition of impact ejecta. The term "peening texture" is suggested, because it is absolutely analogous to the plastic-deformation induced, metal-surface textures generated by ball-bearing bombardment used in engineering metallurgy to work-harden metal surfaces. Additional information is contained in the original (Figures).

Marshall, J. R.↗

Tracking brachytherapy sources using emission imaging with one flat panel detector

This work proposes to use the radiation from brachytherapy sources to track their dwell positions in three-dimensional (3D) space. The prototype device uses a single flat panel detector and a BB tray. The BBs are arranged in a defined pattern. The shadow of the BBs on the flat panel is analyzed to derive the 3D coordinates of the illumination source, i.e., the dwell position of the brachytherapy source. A kilovoltage x-ray source located 3.3 m away was used to align the center BB with the center pixel on the flat panel detector. For a test plan of 11 dwell positions, with an Ir-192 high dose rate unit, one projection was taken for each dwell point, and locations of the BB shadows were manually identified on the projection images. The 3D coordinates for the 11 dwell positions were reconstructed based on two BBs. The distances between dwell points were compared with the expected values. The average difference was 0.07 cm with a standard deviation of 0.15 cm. With automated BB shadow recognition in the future, this technique possesses the potential of tracking the 3D trajectory and the dwell times of a brachytherapy source in real time, enabling real time source position verification.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Building a Foundation for Structure-Based Cellulosome Design for Cellulosic Ethanol: Insight into Cohesin-Dockerin Complexation from Computer Simulation

The organization and assembly of the cellulosome, an extracellular multienzyme complex produced by anaerobic bacteria, is mediated by the high-affinity interaction of cohesin domains from scaffolding proteins with dockerins of cellulosomal enzymes. We have performed molecular dynamics simulations and free energy calculations on both the wild type (WT) and D39N mutant of the C. thermocellum Type I cohesin-dockerin complex in aqueous solution. The D39N mutation has been experimentally demonstrated to disrupt cohesin-dockerin binding. The present MD simulations indicate that the substitution triggers significant protein flexibility and causes a major change of the hydrogen-bonding network in the recognition strips - the conserved loop regions previously proposed to be involved in binding - through electrostatic and salt-bridge interactions between {beta}-strands 3 and 5 of the cohesin and {alpha}-helix 3 of the dockerin. The mutation-induced subtle disturbance in the local hydrogen-bond network is accompanied by conformational rearrangements of the protein side chains and bound water molecules. Additional free energy perturbation calculations of the D39N mutation provide differences in the cohesin-dockerin binding energy, thus offering a direct, quantitative comparison with experiments. The underlying molecular mechanism of cohesin-dockerin complexation is further investigated through the free energy profile, that is, potential of mean force (PMF) calculations of WT cohesin-dockerin complex. The PMF shows a high-free energy barrier against the dissociation and reveals a stepwise pattern involving both the central {beta}-sheet interface and its adjacent solvent-exposed loop/turn regions clustered at both ends of the {beta}-barrel structure.

adaptive biasing force↗

Adapting In Situ Accelerators for Sparsity With Granular Matrix Reordering

Neural network (NN) inference is an essential part of modern systems and is found at the heart of numerous applications ranging from image recognition to natural language processing. In situ NN accelerators can efficiently perform NN inference using resistive crossbars, which makes them a promising solution to the data movement challenges faced by conventional architectures. Although such accelerators demonstrate significant potential for dense NNs, they often do not benefit from sparse NNs, which contain relatively few non-zero weights. Processing sparse NNs on in situ accelerators results in wasted energy to charge the entire crossbar where most elements are zeros. To address this limitation, this paper proposes Granular Matrix Reordering (GMR): a preprocessing technique that enables an energy-efficient computation of sparse NNs on in situ accelerators. GMR reorders the rows and columns of sparse weight matrices to maximize the crossbars' utilization and minimize the total number of crossbars needed to be charged. The reordering process does not rely on sparsity patterns and incurs no accuracy loss. Finally, GMR achieves an average of 28% and up to 34% reduction in energy consumption over seven pruned NNs across four different pruning methods and network architectures.

97 MATHEMATICS AND COMPUTING↗