Browse Articles

A multicenter retrospective study evaluating the effect of proton pump inhibitors on adjuvant tegafur-uracil/leucovorin efficacy for stage II–III colorectal cancer

Transition path and interface sampling of stochastic Schrödinger dynamics

We study rare transitions in Markovian open quantum systems driven with Gaussian noise, applying transition path and interface sampling methods to trajectories generated by stochastic Schrödinger dynamics. Interface and path sampling offer insights into rare event transition mechanisms while simultaneously establishing a quantitative measure of the associated rate constant. Here, we extend their domain to systems described by stochastic Schrödinger equations. As a specific example, we explore a model of quantum Brownian motion in a quartic double well, consisting of a particle coupled to a Caldeira–Leggett oscillator bath, where we note significant departures from the Arrhenius law at low temperatures due to the presence of an anti-Zeno effect.

Effect of Teté dance on lactation session duration in irritable infants in Peru assessed in a randomized controlled trial

Exploring combined spin-labeling approach for structural studies of mRNA in the human ribosome

In this study, we investigated the structural variability of mRNA in the human ribosome by comparing two spin-labeling strategies: one involving an mRNA analog bearing two spin labels attached to the ribose-phosphate backbone and the other placing labels at the nucleotide bases. The use of two strategies of spin labeling of mRNAs allowed us to study for the first time the effect of the structure and location of spin labels on the measured interspin distances in human ribosome complexes. Experiments using dipolar EPR spectroscopy, supported by molecular dynamics calculations, demonstrated that labels introduced at nucleotide bases provide a higher resolution between mRNA conformations in the ribosome mRNA channel, compared to labels introduced at the ribose-phosphate backbone. Although ribose-phosphate labeling turned out to be less informative on its own for studying mRNA conformations in the ribosome than the previously used base labeling, it can find application in other complex studies of the structure of RNAs and their complexes.

Neural network compression for reinforcement learning tasks

A comparison of probabilistic generative frameworks for molecular simulations

Generative artificial intelligence is now a widely used tool in molecular science. Despite the popularity of probabilistic generative models, numerical experiments benchmarking their performance on molecular data are lacking. In this work, we introduce and explain several classes of generative models, broadly sorted into two categories: flow-based models and diffusion models. We select three representative models: neural spline flows, conditional flow matching, and denoising diffusion probabilistic models, and examine their accuracy, computational cost, and generation speed across datasets with tunable dimensionality, complexity, and modal asymmetry. Our findings are varied, with no one framework being the best for all purposes. In a nutshell, (i) neural spline flows do best at capturing mode asymmetry present in low-dimensional data, (ii) conditional flow matching outperforms other models for high-dimensional data with low complexity, and (iii) denoising diffusion probabilistic models appear the best for low-dimensional data with high complexity. Our datasets include a Gaussian mixture model and the dihedral torsion angle distribution of the Aib9 peptide, generated via a molecular dynamics simulation. We hope our taxonomy of probabilistic generative frameworks and numerical results may guide model selection for a wide range of molecular tasks.

Merging synthetic and real embryo data for advanced AI predictions

What have the XH4+ (X = C, Si, Ge) ions in common? An updated summary

A high level quantum mechanical study has been performed to explore the structural rearrangement and relative stability of the XH4+ (X = C, Si, Ge) radical cations at their X̃2T2 ground electronic states. All the stationary points located on the lowest adiabatic sheet of the Jahn–Teller (JT) split X̃2T2 state are fully optimized and characterized by performing harmonic vibrational frequency calculations. Five JT distorted stationary points with D2d(B22), C3v(A12), C2v(B22), and Cs(A′2) symmetries are located on the CH4+ ground state potential energy surface (PES), whereas four such structures are found on each of the SiH4+ and GeH4+ PESs. While the C2v(B22) isomer is found to be a global minimum and the Cs(A′2) one as a transition state for CH4+, the nature of them is reversed for SiH4+ and GeH4+. In particular, the Cs(A′2) stationary points are now global minima for the latter pair of radical cations, and C2v(B22) represents the transition state. Attempts are being made to understand such inconsistent findings via a combination of JT and epikernel principles. The barriers between equivalent C2v(B22) global minimum structures for CH4+ are found to be low, and thus CH4+ undergoes rapid interconversion along cyclic exchange of three hydrogen atoms via Cs transition state. The general features of the ground state PESs of SiH4+ and GeH4+ are similar. The pseudorotation between the Cs lowest energy structures undergoes along SiH2 and GeH2 wagging motions via C2v(B22) transition state for SiH4+ and GeH4+, respectively.

Integrated multi-omics unveils novel immune signature for predicting prognosis in colon cancer patients

Adaptive-precision potentials for large-scale atomistic simulations

Large-scale atomistic simulations rely on interatomic potentials, providing an efficient representation of atomic energies and forces. Modern machine-learning (ML) potentials provide the most precise representation compared to electronic structure calculations, while traditional potentials provide a less precise but computationally much faster representation and, thus, allow simulations of larger systems. We present a method to combine a traditional and a ML potential into a multi-resolution description, leading to an adaptive-precision potential with an optimum of performance and precision in large, complex atomistic systems. The required precision is determined per atom by a local structure analysis and updated automatically during simulation. We use copper as demonstrator material with an embedded atom model as classical force field and an atomic cluster expansion (ACE) as ML potential, but, in principle, a broader class of potential combinations can be coupled by this method. The approach is developed for the molecular-dynamics simulator LAMMPS and includes a load-balancer to prevent problems due to the atom dependent force-calculation times, which makes it suitable for large-scale atomistic simulations. The developed adaptive-precision copper potential represents the ACE-forces with a precision of 10 me V/Å and the ACE-energy exactly for the precisely calculated atoms in a nanoindentation of 4 × 106 atoms calculated for 100 ps and shows a speedup of 11.3 compared with a full ACE simulation.

Optical coherence tomography angiography compared with optical coherence tomography for detection of glaucoma progression with high myopia

Molecular structures with spectroscopic accuracy at DFT cost by the templating synthon approach and the PCS141 database

The computation of accurate geometric parameters at density functional theory cost for large molecules in the gas phase is addressed through a novel strategy that combines quantum chemical models with machine learning techniques. The first key step is the expansion of a database of accurate semi-experimental equilibrium structures with additional molecular geometries optimized by version 2 of the Pisa composite scheme. Then, the templating synthon approach is used to improve the accuracy of structures optimized by a hybrid density functional paired with a double zeta basis set, leveraging chemical similarity to cluster different molecular environments and refine bond lengths and valence angles. A set of prototypical biomolecular building blocks is used to demonstrate that it is possible to achieve spectroscopic accuracy for molecular systems too large to be treated by state-of-the-art composite wavefunction methods. In addition, a freely accessible web-based tool has been developed to facilitate the post-processing of geometries optimized using standard electronic structure codes, thereby providing an accurate and efficient tool for the computational study of medium- to large-sized molecules, also accessible to experiment-oriented researchers.

Optimizing data privacy and security measures for critical infrastructures via IoT based ADP2S technique

How the choice of exchange–correlation functional affects DFT-based simulations of the hydrated electron

Hydrated electrons are anionic species that are formed when an excess electron is introduced into liquid water. Building an understanding of how hydrated electrons behave in solution has been a long-standing effort of simulation methods, of which density functional theory (DFT) has come to the fore in recent years. The ability of DFT to model the reactive chemistry of hydrated electrons is an attractive advantage over semi-classical methodologies; however, relatively few density functional approximations (DFAs) have been used for the hydrated electron simulations presented in the literature. Here, we simulate hydrated electron systems using a series of exchange–correlation (XC) functionals spanning Jacob’s ladder. We calculate a variety of experimental and other observables of the hydrated electron and compare the XC functional dependence for each quantity. We find that the formation of a stable localized hydrated electron is not necessarily limited to hybrid XC functionals and that some hybrid functionals produce delocalized hydrated electrons or electrons that react with the surrounding water at an unphysically fast rate. We further characterize how different DFAs impact the solvent structure and predicted spectroscopy of the hydrated electron, considering several methods for calculating the hydrated electron’s absorption spectrum for the best comparison between structures generated using different density functionals. None of the dozen or so DFAs that we investigated are able to correctly predict the hydrated electron’s spectroscopy, vertical detachment energy, or molar solvation volume.

Adaptive hybrid segmentation combined with meta heuristic optimization in transfer learning for plant leaf disease classification

Emergent marginality in frustrated multistable networks

We study disordered networks of coupled bistable elastic elements, representing a coarse-grained view of amorphous solids. We find that such networks self-organize to a marginally stable state, in which the barrier for local activations becomes vanishingly small. The model provides unique access to both local and global properties associated with marginal stability. We directly measure pseudo-gaps in the spectrum of local excitations, as well as diverging fluctuations under shear. Crucially, the dynamics are dominated by a small population of bonds that are locally unstable, which give rise to quasi-localized, low-frequency vibrational modes and scale-free avalanches of instabilities. We propose a correction to the scaling between the pseudo-gap exponent and avalanche statistics based on diverging length fluctuations. Our model combines a coarse-grained view with a continuous, real-space implementation, providing novel insights to a wide class of amorphous solids.

HD-tDCS effects on social impairment in autism spectrum disorder with sensory processing abnormalities: a randomized controlled trial

Perturbation of water–ethanol solvent structural relaxation by a bis-urea supramolecular gel and paracetamol

Understanding structural dynamics on the picosecond/nanometer scale in complex fluids is crucial for advancing various fields, from material chemistry to drug delivery. We employ polarized quasi-elastic neutron spectroscopy to investigate the perturbation to the hydrogen bond network of water–ethanol mixtures induced by a supramolecular gel network and by paracetamol (PCM) molecules. Interestingly, while the supramolecular gelator significantly alters the macroscopic behavior of the solvent at concentrations of 0.3 and 0.5 wt. %, it does not affect the hydrogen bond network at the microscopic level. In contrast, the addition of PCM at 5 wt. %, which does not change the macroscopic properties, modifies the structural dynamics of water–ethanol mixtures at length scales commensurate with and below the PCM–PCM correlation length in the mixture. This study reveals the intricate interplay between solute, solvent, and gel interactions, demonstrating a lack of direct correlation between macroscopic and microscopic properties in such complex systems.

Biofilm proficient Bacillus subtilis prevents neurodegeneration in Caenorhabditis elegans Parkinson’s disease models via PMK-1/p38 MAPK and SKN-1/Nrf2 signaling

Effect of structural bending on the photophysical properties of perylene bisimide

The effect of nonplanarity on the electronic properties of π-systems has been difficult to study systematically because of the limited availability of suitable model compounds. Our group recently synthesized a series of end-to-end bent perylene bisimide (PBI) cyclophanes, whose degree of bending is adjustable by modifying the internal alkyl tethers. Herein, we subjected these bent PBI derivatives to theoretical calculations and time-resolved spectroscopy. The current study has offered rational explanations for several unique photophysical characteristics of bent PBIs: (1) the redshifts of the S0–S1 transitions, (2) the decrease in extinction coefficients, (3) the broadening of spectral shapes, and (4) the suppression of nonradiative decay processes. Furthermore, the investigation of the S1 states and radical anions has revealed that structural bending also substantially alters the energy levels of upper molecular orbitals such as LUMO+2.