Machine learning for detection of 3D features using sparse x-ray tomographic reconstruction.

In many inertial confinement fusion (ICF) experiments, the neutron yield and other parameters cannot be completely accounted for with one and two dimensional models. This discrepancy suggests that there are three dimensional effects that may be significant. Sources of these effects include defects in the shells and defects in shell interfaces, the fill tube of the capsule, and the joint feature in double shell targets. Due to their ability to penetrate materials, x rays are used to capture the internal structure of objects. Methods such as computational tomography use x-ray radiographs from hundreds of projections, in order to reconstruct a three dimensional model of the object. In experimental environments, such as the National Ignition Facility and Omega-60, the availability of these views is scarce, and in many cases only consists of a single line of sight. Mathematical reconstruction of a 3D object from sparse views is an ill-posed inverse problem. These types of problems are typically solved by utilizing prior information. Neural networks have been used for the task of 3D reconstruction as they are capable of encoding and leveraging this prior information. We utilize half a dozen, different convolutional neural networks to produce different 3D representations of ICF implosions from the experimental data. Deep supervision is utilized to train a neural network to produce high-resolution reconstructions. These representations are used to track 3D features of the capsules, such as the ablator, inner shell, and the joint between shell hemispheres. Machine learning, supplemented by different priors, is a promising method for 3D reconstructions in ICF and x-ray radiography, in general.

Aquaglyceroporin AQP7's affinity for its substrate glycerol: Have we reached convergence in the computed values of glycerol-aquaglyceroporin affinity?

AQP7 is one of the four human aquaglyceroporins that facilitate glycerol transport across the cell membrane, a biophysical process that is essential in human physiology. Therefore, it is interesting to compute AQP7's affinity for its substrate (glycerol) with reasonable certainty to compare with the experimental data suggesting high affinity in contrast with most computational studies predicting low affinity. In this study aimed at computing the AQP7-glycerol affinity with high confidence, we implemented a direct computation of the affinity from unbiased equilibrium molecular dynamics (MD) simulations of three all-atom systems constituted with 0.16M, 4.32M, and 10.23M atoms, respectively. These three sets of simulations manifested a fundamental physics law that the intrinsic fluctuations of pressure in a system are inversely proportional to the system size (the number of atoms in it). These simulations showed that the computed values of glycerol-AQP7 affinity are dependent upon the system size (the inverse affinity estimations were, respectively, 47.3 mM, 1.6 mM, and 0.92 mM for the three model systems). In this, we obtained a lower bound for the AQP7-glycerol affinity (an upper bound for the dissociation constant). Namely, the AQP7-glycerol affinity is stronger than 1087/M (the dissociation constant is less than 0.92 mM). Additionally, we conducted hyper steered MD (hSMD) simulations to map out the Gibbs free-energy profile. From the free-energy profile, we produced an independent computation of the AQP7-glycerol dissociation constant being approximately 0.18 mM.

In silico simulations of erythrocyte aquaporins with quantitative in vitro validation.

Modelling water and membrane lipids is an essential element in the computational research of biophysical/biochemical processes such as water transport across the cell membrane. In this study, we examined the accuracies of two popular water models, TIP3P and TIP4P, in the molecular dynamics simulations of erythrocyte aquaporins (AQP1 and AQP3). We modelled the erythrocyte membrane as an asymmetric lipid bilayer with appropriate lipid compositions of its inner and outer leaflet, in comparison with a symmetric lipid bilayer of a single lipid type. We computed the AQP1/3 permeabilities with the transition state theory with full correction for recrossing events. We also conducted cell swelling assays for water transport across the erythrocyte membrane. The experimental results agree with the TIP3P water-erythrocyte membrane model, in confirmation of the expected accuracy of the erythrocyte membrane model, the TIP3P water model, and the CHARMM parameters for water-protein interactions.