Search

Motivation Characterizing drug–protein interactions (DPIs) is crucial to the high-throughput screening for drug discovery. The deep learning-based approaches have attracted attention because they can predict DPIs without human trial and error. However, because data labeling requires significant resources, the available protein data size is relatively small, which consequently decreases model performance. Here, we propose two methods to construct a deep learning framework that exhibits superior performance with a small labeled dataset. Results At first, we use transfer learning in encoding protein sequences with a pretrained model, which trains general sequence representations in an unsupervised manner. Second, we use a Bayesian neural network to make a robust model by estimating the data uncertainty. Our resulting model performs better than the previous baselines at predicting interactions between molecules and proteins. We also show that the quantified uncertainty from the Bayesian inference is related to confidence and can be used for screening DPI data points. Availability and implementation The code is available at https://github.com/QHwan/PretrainDPI. Supplementary information Supplementary data are available at Bioinformatics online. [ABSTRACT FROM AUTHOR]

Ionic liquids (ILs) are emerging as novel solvents that exhibit peculiar mechanical properties in the form of thin films on metal surfaces under normal pressure. However, the mechanical properties of ILs in the form of nano-meniscus have not been analyzed yet. Here, we investigate the shear viscoelasticity of a single IL meniscus at the nanoscale. To characterize the shear rheological properties of ILs, we employ a quartz tuning fork-based atomic force microscope, conduct dynamic force spectroscopy, and analyse shear properties using the non-Newtonian–Maxwell model. The elastic response of the IL nanomeniscus is found to be about 25 times higher than that of the bulk IL bridge, whereas the viscous responses are similar. In addition, by conducting shear velocity-dependent measurements, we find that the IL meniscus shows nonlinear rheological behaviours. Interestingly, we observe that the relaxation time of the IL increases at a tip–substrate distance of about 60 nm. [ABSTRACT FROM AUTHOR]

Understanding the phases of water molecules based on local structure is essential for understanding their anomalous properties. However, due to complicated structural motifs formed via hydrogen bonds, conventional order parameters represent water molecules incompletely. In this paper, we develop GCIceNet, which automatically generates machine-based order parameters for classifying the phases of water molecules via supervised and unsupervised learning. The multiple graph convolutional layers in GCIceNet can learn topological information on the complex hydrogen bond networks. It shows a substantial improvement in accuracy for predicting the phase of water molecules in a bulk system and an ice/vapor interface system. A relative importance analysis shows that GCIceNet can capture the structural features of the given system hidden in the input data. Augmented with the vast amount of data provided by molecular dynamics simulations, GCIceNet is expected to serve as a powerful tool for the fields of glassy liquids and hydration layers around biomolecules. [ABSTRACT FROM AUTHOR]

We present control of force sensitivity in Q-controlled amplitude-modulation atomic force microscopy (AM-AFM) that is based on the high-Q quartz tuning-fork. It is found that the phase noise is identical to the amplitude noise divided by oscillation amplitude in AM-AFM. In particular, we observe that while Q-control does not compromise the signal-to-noise ratio, it enhances the detection sensitivity because the minimum detectable force gradient is inversely proportional to the effective quality factor for large bandwidths, which is due to reduction of frequency noise. This work demonstrates Q-control in AM-AFM is a useful technique for enhancement of the force sensitivity with increased Q or improvement of the scanning speed with decreased Q. [ABSTRACT FROM AUTHOR]

Q-control technique enables to actively change the quality factor of the probe oscillation in dynamic atomic force microscopy. The Q-control is realized by adding a self-feedback loop into the original actuation-detection system, in which a damping force with controllable damping coefficient in magnitude and sign is applied to the oscillating probe. While the applied force alters the total damping interaction and thus the overall ‘signal’ of the probe motion, the added feedback system changes the ‘noise’ of the motion as well. Here, we systematically investigate the signal, the noise, and the signal-to-noise ratio of the qPlus sensor under the active Q-control. We quantify the noise of the qPlus motion by measuring the noise spectral density, which is reproduced by a harmonic oscillator model including the thermal and the measurement noises. We show that the noise signal increases with the quality factor controlled, scaling as the square root of the quality factor. Because the overall signal is linearly proportional to the quality factor, the signal-to-noise ratio scales as the square root of the quality factor. The Q-controlled qPlus with a highly enhanced Q, up to 10,000 in air, leads to the minimum detectable force gradient of 0.001 N/m, which would enhance the capability of the qPlus sensor for atomic force microscopy and spectroscopy. [ABSTRACT FROM AUTHOR]

The viscometry of minute amounts of liquid has been in high demand as a novel tool for medical diagnosis and biological assays. Various microrheological techniques have shown the capability to handle small volumes. However, as the liquid volume decreases down to nanoliter scale, increasingly dominant surface effects complicate the measurement and analysis, which remain a challenge in microrheology. Here, we demonstrate an atomic force microscope-based platform that determines the viscosity of single sessile drops of 1 nanoliter Newtonian fluids. We circumvent interfacial effects by measuring the negative-valued shear elasticity, originating from the retarded fluidic response inside the drop. Our measurement is independent of the liquid-boundary effects, and thus is valid without a priori knowledge of surface tension or contact angle, and consistently holds at a 1 milliliter-scale volume. Importantly, while previous methods typically need a much larger ‘unrecoverable’ volume above 1 microliter, our simple platform uses only ∼1 nanoliter. Our results offer a quantitative and unambiguous methodology for viscosity measurements of extremely minute volumes of Newtonian liquids on the nanoliter scale. [ABSTRACT FROM AUTHOR]

Our ability to control friction remains modest, as our understanding of the underlying microscopic processes is incomplete. Atomic force experiments have provided a wealth of results on the dependence of nanofriction on structure, velocity and temperature, but limitations in the dynamic range, time resolution, and control at the single-atom level have hampered a description from first principles. Here, using an ion-crystal system with single-atom, single-substrate-site spatial and single-slip temporal resolution, we measure the friction force over nearly five orders of magnitude in velocity, and contiguously observe four distinct regimes, while controlling temperature and dissipation. We elucidate the interplay between thermal and structural lubricity for two coupled atoms, and provide a simple explanation in terms of the Peierls-Nabarro potential. This extensive control at the atomic scale enables fundamental studies of the interaction of many-atom surfaces, possibly into the quantum regime. [ABSTRACT FROM AUTHOR]

We demonstrate real-time observation of nucleation of the single water nanomeniscus formed via capillary condensation. We directly measure (i) activation time by time-resolved atomic force microscopy and (ii) nucleation rate by statistical analysis of its exponential distribution, which is the experimental evidence that the activation process is stochastic and follows the Poisson statistics. It implies that formation of the water nanomeniscus is triggered by nucleation, which requires activation for producing a nucleus. We also find the dependence of the nucleation rate on the tip-sample distance and temperature. [ABSTRACT FROM AUTHOR]

We present an experimental observation and comprehensive theory that address the effect of interaction on the phase transition in a many-body system, which is far from equilibrium and does not have detailed balance. In the resonantly driven cold atomic trap, we observe kinetic phase transition (KPT) that is manifested by substantial enhancement of fluctuations, similar to the first-order phase transition in thermal equilibrium. Moreover, we can control the attractive interaction between atoms trapped in the coexisting states by adjusting the total number of atoms, which induces shift of the phase boundaries of KPT. The demonstrated effect is the many-body analogue of KPT that was previously predicted and observed in a single-driven oscillator. Our work provides a unique model platform for quantitative study of nonequilibrium nonlinear dynamics in many-body systems. [ABSTRACT FROM AUTHOR]

The mechanical properties (viscoelasticity) of the nanoscale molecular cluster of water meniscus, spontaneously formed between a quartz tip (∼100 nm curvature) and a mica substrate were quantitatively studied. The theoretical and experimental investigation was performed on the basis of the quartz tuning fork-based frequency modulation-atomic force microscope system with a high vertical resolution (∼0.5 Å). The proposed system is suitable apparatus for the dynamic force spectroscopy of nanoscopic materials with several advantages including high sensitivity, short response time, immunity to the electrical noise, and simple and intuitive interpretation of the results using the frequency shift. [ABSTRACT FROM AUTHOR]

We report on our experimental study of spectral diffusion in grown InGaAs single quantum dots. Using a double-tapered tip, near-field scanning optical microscope, we obtain the spectral diffusion resulting from the quantum confined Stark effect of individual quantum dots in a randomly fluctuating electric field. These electric fields come from localized charges at defects in the vicinity of the quantum dots. In particular, we find that the similar-patterned temporal jitter, seen in the spectral diffusion, emits from a quantum dot and is dependent on the excitation density and applied bias voltage. [ABSTRACT FROM AUTHOR]