Prediction of subcellular localization using sequence-biased recurrent networks

Motivation: Targeting peptides direct nascent proteins to their specific subcellular compartment. Knowledge of targeting signals enables informed drug design and reliable annotation of gene products. However, due to the low similarity of such sequences and the dynamical nature of the sorting process, the computational prediction of subcellular localization of proteins is challenging. Results: We contrast the use of feed forward models as employed by the popular TargetP/SignalP predictors with a sequence-biased recurrent network model. The models are evaluated in terms of performance at the residue level and at the sequence level, and demonstrate that recurrent networks improve the overall prediction performance. Compared to the original results reported for TargetP, an ensemble of the tested models increases the accuracy by 6 and 5% on non-plant and plant data, respectively.

Effect of pore packing defects in 2-D ordered mesoporous carbons on ionic transport

Ordered mesoporous materials show great importance in energy, environmental, and chemical engineering. The diffusion of guest species in mesoporous networks plays an important role in these applications, especially for energy storage, such as supercapacitors based on ordered mesoporous carbons ( OMCs). The ion diffusion behavior in two different 2-D hexagonal OMCs was investigated by using cyclic voltametry and electrochemical impedance spectroscopy. In addition, transmission electron microscopy, small-angle X-ray diffraction, and nitrogen cryosorption methods were used to study the pore structure variations of these two OMCs. It was found that, for the OMC with defective pore channels ( termed as pore packing defects), the gravimetric capacitance was greatly decayed when the voltage scan rate was increased. The experimental results suggest that, for the ion diffusion in 2-D hexagonal OMCs with similar mesopore size distribution, the pore packing defect is a dominant dynamic factor.