ProtInteract: A deep learning framework for predicting protein-protein interactions.

Proteins mainly perform their functions by interacting with other proteins. Protein-protein interactions underpin various biological activities such as metabolic cycles, signal transduction, and immune response. However, due to the sheer number of proteins, experimental methods for finding interacting and non-interacting protein pairs are time-consuming and costly. We therefore developed the ProtInteract framework to predict protein-protein interaction. ProtInteract comprises two components: first, a novel autoencoder architecture that encodes each protein's primary structure to a lower-dimensional vector while preserving its underlying sequence attributes. This leads to faster training of the second network, a deep convolutional neural network (CNN) that receives encoded proteins and predicts their interaction under three different scenarios. In each scenario, the deep CNN predicts the class of a given encoded protein pair. Each class indicates different ranges of confidence scores corresponding to the probability of whether a predicted interaction occurs or not. The proposed framework features significantly low computational complexity and relatively fast response. The contributions of this work are twofold. First, ProtInteract assimilates the protein's primary structure into a pseudo-time series. Therefore, we leverage the nature of the time series of proteins and their physicochemical properties to encode a protein's amino acid sequence into a lower-dimensional vector space. This approach enables extracting highly informative sequence attributes while reducing computational complexity. Second, the ProtInteract framework utilises this information to identify protein interactions with other proteins based on its amino acid configuration. Our results suggest that the proposed framework performs with high accuracy and efficiency in predicting protein-protein interactions.

Protein-protein interaction prediction with deep learning: A comprehensive review.

Most proteins perform their biological function by interacting with themselves or other molecules. Thus, one may obtain biological insights into protein functions, disease prevalence, and therapy development by identifying protein-protein interactions (PPI). However, finding the interacting and non-interacting protein pairs through experimental approaches is labour-intensive and time-consuming, owing to the variety of proteins. Hence, protein-protein interaction and protein-ligand binding problems have drawn attention in the fields of bioinformatics and computer-aided drug discovery. Deep learning methods paved the way for scientists to predict the 3-D structure of proteins from genomes, predict the functions and attributes of a protein, and modify and design new proteins to provide desired functions. This review focuses on recent deep learning methods applied to problems including predicting protein functions, protein-protein interaction and their sites, protein-ligand binding, and protein design.

Eighty-Five Percent of Improved Optical Power Delivery to Epiretinal Prostheses Using Rigid Body Compensation Algorithm.

Vision impairment caused by degenerative retinal pathologies such as age-related macular degeneration can be treated using retinal implants. Such devices receive power and data using cables passing through a permanent surgical incision in the eye wall (sclera), which increases the risk to patients and surgical costs. A recently developed retinal implant design eliminates the necessity of the implant cable using a photonic power converter (PPC), which receives optical power and data through the pupil and is directed by an ellipsoidal reflector and micro-electromechanical mirror. We present a misalignment compensation algorithm model that accounts for rigid-body motions of the reflector relative to the eye and applies the correction to the mirror coordinates in the presence of angular misalignment of the reflector. We demonstrate that up to 85% of the nominal optical power can be delivered to the implant with axial reflector misalignments up to 30 deg using the compensation algorithm.

The Implication of Spatial Statistics in Human Mesenchymal Stem Cell Response to Nanotubular Architectures.

INTRODUCTION: In recent years there has been ample interest in nanoscale modifications of synthetic biomaterials to understand fundamental aspects of cell-surface interactions towards improved biological outcomes. In this study, we aimed at closing in on the effects of nanotubular TiO2 surfaces with variable nanotopography on the response on human mesenchymal stem cells (hMSCs). Although the influence of TiO2 nanotubes on the cellular response, and in particular on hMSC activity, has already been addressed in the past, previous studies overlooked critical morphological, structural and physical aspects that go beyond the simple nanotube diameter, such as spatial statistics. METHODS: To bridge this gap, we implemented an extensive characterization of nanotubular surfaces generated by anodization of titanium with a focus on spatial structural variables including eccentricity, nearest neighbour distance (NND) and Voronoi entropy, and associated them to the hMSC response. In addition, we assessed the biological potential of a two-tiered honeycomb nanoarchitecture, which allowed the detection of combinatory effects that this hierarchical structure has on stem cells with respect to conventional nanotubular designs. We have combined experimental techniques, ranging from Scanning Electron (SEM) and Atomic Force (AFM) microscopy to Raman spectroscopy, with computational simulations to characterize and model nanotubular surfaces. We evaluated the cell response at 6 hrs, 1 and 2 days by fluorescence microscopy, as well as bone mineral deposition by Raman spectroscopy, demonstrating substrate-induced differential biological cueing at both the short- and long-term. RESULTS: Our work demonstrates that the nanotube diameter is not sufficient to comprehensively characterize nanotubular surfaces and equally important parameters, such as eccentricity and wall thickness, ought to be included since they all contribute to the overall spatial disorder which, in turn, dictates the overall bioactive potential. We have also demonstrated that nanotubular surfaces affect the quality of bone mineral deposited by differentiated stem cells. Lastly, we closed in on the integrated effects exerted by the superimposition of two dissimilar nanotubular arrays in the honeycomb architecture. DISCUSSION: This work delineates a novel approach for the characterization of TiO2 nanotubes which supports the incorporation of critical spatial structural aspects that have been overlooked in previous research. This is a crucial aspect to interpret cellular behaviour on nanotubular substrates. Consequently, we anticipate that this strategy will contribute to the unification of studies focused on the use of such powerful nanostructured surfaces not only for biomedical applications but also in other technology fields, such as catalysis.

Effects of van der Waals Force and Thermal Stresses on Pull-in Instability of Clamped Rectangular Microplates.

We study the influence of von Karman nonlinearity, van der Waals force, and a athermal stresses on pull-in instability and small vibrations of electrostatically actuated mi-croplates. We use the Galerkin method to develop a tractable reduced-order model for elec-trostatically actuated clamped rectangular microplates in the presence of van der Waals forcesand thermal stresses. More specifically, we reduce the governing two-dimensional nonlineartransient boundary-value problem to a single nonlinear ordinary differential equation. For thestatic problem, the pull-in voltage and the pull-in displacement are determined by solving apair of nonlinear algebraic equations. The fundamental vibration frequency corresponding toa deflected configuration of the microplate is determined by solving a linear algebraic equa-tion. The proposed reduced-order model allows for accurately estimating the combined effectsof van der Waals force and thermal stresses on the pull-in voltage and the pull-in deflectionprofile with an extremely limited computational effort.