Comparison of Machine Learning Methods towards Developing Interpretable Polyamide Property Prediction.

Polyamides are often used for their superior thermal, mechanical, and chemical properties. They form a diverse set of materials that have a large variation in properties between linear to aromatic compounds, which renders the traditional quantitative structure-property relationship (QSPR) challenging. We use extended connectivity fingerprints (ECFP) and traditional QSPR fingerprints to develop machine learning models to perform high fidelity prediction of glass transition temperature (Tg), melting temperature (Tm), density (ρ), and tensile modulus (E). The non-linear model using random forest is in general found to be more accurate than linear regression; however, using feature selection or regularization, the accuracy of linear models is shown to be improved significantly to become comparable to the more complex nonlinear algorithm. We find that none of the models or fingerprints were able to accurately predict the tensile modulus E, which we hypothesize is due to heterogeneity in data and data sources, as well as inherent challenges in measuring it. Finally, QSPR models revealed that the fraction of rotatable bonds, and the rotational degree of freedom affects polyamide properties most profoundly and can be used for back of the envelope calculations for a quick estimate of the polymer attributes (glass transition temperature, melting temperature, and density). These QSPR models, although having slightly lower prediction accuracy, show the most promise for the polymer chemist seeking to develop an intuition of ways to modify the chemistry to enhance specific attributes.

Improving the torsion performance of IPMC by changing the electrode separation.

Ionic polymer metal composites (IPMCs) are widely studied as actuators and sensors, due to their large bending motion, flexibility and being light-weight. Nowadays, IPMCs are used in the bionic field, for example, to achieve bending and twisting movements of wings and fins. In this paper, a method is proposed to optimize the torsion performance of IPMCs by changing the electrode separation. The IPMCs with patterned electrode fabricated by masking technique are proposed to accomplish twisting motion. The result indicates that the torsion performance is improved as the electrode separation increased. Thereby it provides a new strategy for the bionic field with twisting behavior.

The microscale Weissenberg effect for high-viscosity solution pumping at the picoliter level.

Transportation of highly viscous solutions at the picoliter level with a rapid dynamic response is paramount for micro/nano-fabrication. With the advantages of a higher length-wall (thickness) ratio and a more stable free surface compared to those of the traditional Weissenberg effect (TWE), the microscale Weissenberg effect (MWE) can continuously and controllably pump high-viscosity solutions at the picoliter scale. Some typical characteristics and behaviors of MWE are investigated as the rotation rod diameter decreases to the microscale of ∼100 µm. The pumped minimum solution volume can reach 167.5 pL per second, and the minimum response time of solution pumping is 0.3 s, which is much shorter than that of pressure driven pumping. Then, a new direct writing with an adjustable jet diameter based on the MWE is proposed to write microstructures on a substrate from a solution with a viscosity of approximately 130.1 Pa s. The stability of the as-spun jet and the deposited structures is improved when a high voltage is applied. To fully demonstrate the advantages of MWE, MWE-based direct writing is performed to successfully fabricate microfluidic channels with variable diameters. Thus, the system can overcome the problems of high transport resistance to the pumping of a high-viscosity solution.

Single Nucleotide Polymorphisms in SLC19A1 and SLC25A9 Are Associated with Childhood Autism Spectrum Disorder in the Chinese Han Population.

Genetic variants have been implicated in the development of autism spectrum disorder (ASD). Recent studies suggest that solute carriers (SLCs) may play a role in the etiology of ASD. This purpose of this study was to determine the association between single nucleotide polymorphisms (SNPs) in SLC19A1 and SLC25A12 genes with childhood ASD in a Chinese Han population. A total of 201 autistic children and 200 age- and gender-matched healthy controls were recruited. A TaqMan probe-based real-time PCR approach was used to determine genotypes of SNPs corresponding to rs1023159 and rs1051266 in SLC19A1, and rs2056202 and rs2292813 in SLC25A12. Our results showed that both the T/T genotype of rs1051266 (odds ratio (OR) = 1.85, 95% confidence interval (CI) = 1.06-3.23, P = 0.0301) and the T allele (OR = 1.77, 95% CI = 1.07-2.90, P = 0.0249) of rs2292813 were significantly associated with an increased risk of childhood ASD. In addition, the G-C haplotype of rs1023159-rs1051266 in SCL19A1 (OR = 0.71, 95% CI = 0.51-0.98, P = 0.0389) and C-C haplotype of rs2056202-rs2292813 in SLC25A12 (OR = 0.58, 95% CI = 0.35-0.96, P = 0.0325) were associated with decreased risks of childhood ASD. There was no significant association between genotypes and allele frequencies with the severity of the disease. Our study suggests that these genetic variants of SLC19A1 and SLC25A12 may be associated with risks for childhood ASD.

The fast fabrication of flexible electronic devices of graphene composites.

The rapid production and accurate deposition of graphene composites are first integrated into one process, due to the strong interaction between the polymer bond with sodium dodecyl sulfonate (SDS) and graphene. It is demonstrated that tension-shear exfoliation in high viscosity fluid may get a higher graphene production rate than in N-methyl-pyrrolidone. In addition, the micro-scale patterns of graphene nanomaterials produced by this method show high electrical conductivity and superior sensitivity to pressure due to their porous structure.

Potentials of mean force and escape times of surfactants from micelles and hydrophobic surfaces using molecular dynamics simulations.

We calculate potentials of mean force (PMFs) and mean first passage times for a surfactant to escape a micelle, for both ionic sodium dodecyl sulfate (SDS) and nonionic ethoxylated alcohol (C12E5) micelles using both atomistic and coarse-grained molecular dynamics (MD) simulations. The PMFs are obtained by umbrella sampling and used in a Smoluchowski first-passage-time theory to obtain the times for a surfactant to escape a micelle. The calculated mean first passage time for an SDS molecule to break away from a micelle (with an aggregation number of 60) is around 2 µs, which is consistent with previous experimental measurements of the "fast relaxation time" for exchange of surfactants between the micellar phase and the bulk solvent. The corresponding escape time calculated for a nonionic ethoxylated alcohol C12E5, with the same tail length as SDS, is 60 µs, which is significantly longer than for SDS primarily because the PMF for surfactant desorption is about 3kT smaller than for C12E5. We also show that two coarse-grained (CG) force fields, MARTINI and SDK, give predictions similar to the atomistic CHARMM force field for the nonionic C12E5 surfactant, but for the ionic SDS surfactant, the CG simulations give a PMF similar to that obtained with CHARMM only if long-range electrostatic interactions are included in the CG simulations, rather than using a shifted truncated electrostatic interaction. We also calculate that the mean first passage time for an SDS and a C12E5 to escape from a latex binder surface is of the order of milliseconds, which is more than 100 times longer than the time for escape from the micelle, indicating that in latex waterborne coatings, SDS and C12E5 surfactants likely bind preferentially to the latex polymer interface rather than form micelles, at least at low surfactant concentrations.

Coarse-grained molecular dynamics simulation of self-assembly and surface adsorption of ionic surfactants using an implicit water model.

We perform coarse-grained molecular dynamics simulations for sodium dodecyl sulfate (SDS) surfactant using a modification of the Dry Martini force field (Arnarez et al. 2014) with implicit water. After inclusion of particle mesh Ewald (PME) electrostatics, an artificially high dielectric constant for water (ε(r) = 150), and reparameterization, we obtain structural and thermodynamic properties of SDS micelles that are close to those obtained from the standard Martini force field with explicit water, which in turn match those of atomistic simulations. The gains in computational efficiency obtained by removing explicit water allow direct simulations of the self-assembly of SDS in solution. We observe surfactant exchange among micelles and micelle fission and fusion and obtain realistic, equilibrated micelle size distributions at modest computational cost, as well as a transition to cylindrical micelles at high surfactant concentration or with added salt. We further apply this parametrized force field to study the adsorption of SDS onto hydrophobic surfaces and calculate the adsorption kinetics and equilibrium adsorption isotherm. The greatly increased speed of computation of surfactant self-assembly made possible by this Dry Martini method should allow future simulation of competitive adsorption of multiple surfactant species to surfaces, as well as simulation of micellar shape transitions.

Water channel formation and ion transport in linear and branched lipid bilayers.

Using molecular dynamics simulations, we studied the influence of methyl chain branching on transmembrane potential induced formation of water channels in lipid bilayers and ion transport. We compared the response of a bilayer lipid that has multiple methyl branches diphytanoylphosphatidylcholine (DPhPC) with its straight-chain counterpart dipalmitoylphosphatidylcholine (DPPC) to a transmembrane potential created by an imbalance in ionic charges across the membrane. We found that, compared to the straight-chain DPPC lipid bilayer membranes, branched DPhPC lipid membranes require a higher critical transmembrane potential to break down, followed by water channel formation, and transport of anions and cations through the pore. We demonstrated that the bulkiness of the added methyl branches leads to "barrel-stave" pores in DPhPC membranes which require a higher transmembrane potential to produce than the toroidal pores produced in the straight chain DPPC lipid bilayers. Our results provided a deeper understanding of the water channel formation and ion transport through lipid bilayer membrane and might help explain the increased resistance to charge-induced poration in organisms with membranes abundant in branched lipids.

Proteins searching for their target on DNA by one-dimensional diffusion: overcoming the "speed-stability" paradox.

The sequence dependence of DNA-protein interactions that allows proteins to find the correct reaction site also slows down the 1D diffusion of the protein along the DNA molecule, leading to the so-called "speed-stability paradox," wherein fast diffusion along the DNA molecule is seemingly incompatible with stable targeting of the reaction site. Here, we develop diffusion-reaction models that use discrete and continuous Gaussian random 1D diffusion landscapes with or without a high-energy cut-off, and two-state models with a transition to and from a "searching" mode in which the protein diffuses rapidly without recognizing the target. We show the conditions under which such considerations lead to a predicted speed-up of the targeting process, and under which the presence of a "searching" mode in a two-state model is nearly equivalent to the existence of a high-energy cut-off in a one-state model. We also determine the conditions under which the search is either diffusion-limited or reaction-limited, and develop quantitative expressions for the rate of successful targeting as a function of the site-specific reaction rate, the roughness of the DNA-protein interaction potential, and the presence of a "searching" mode. In general, we find that a rough landscape is compatible with a fast search if the highest energy barriers can be avoided by "hopping" or by the protein transitioning to a lower-energy "searching" mode. We validate these predictions with the results of Brownian dynamics, kinetic Metropolis, and kinetic Monte Carlo simulations of the diffusion and targeting process, and apply these concepts to the case of T7 RNA polymerase searching for its target site on T7 DNA.

Selectivity of ligand-receptor interactions between nanoparticle and cell surfaces.

Selectivity of interactions between nanoparticles functionalized by tethered ligands and cell surfaces with different densities of receptors plays an essential role in biorecognition and its implementation in nanobiomedicine. We show that the onset of nanoparticle adsorption has a universal character for a range of nanoparticles: the onset receptor density decreases exponentially with the energy of ligand-receptor binding and inversely with the ligand density. We demonstrate that a bimodal tether distribution, which permits shielding ligands by longer nonfunctional tethers, leads to extra loss of entropy at the adsorption onset, enhancing the selectivity.

Nanoparticle design optimization for enhanced targeting: Monte Carlo simulations.

Using computer simulations, we systematically studied the influence of different design parameters of a spherical nanoparticle tethered with monovalent ligands on its efficiency of targeting planar cell surfaces containing mobile receptors. We investigate how the nanoparticle affinity can be affected by changing the binding energy, the percent of functionalization by ligands, tether length, grafting density, and nanoparticle core size. In general, using a longer tether length or increasing the number of tethered chains without increasing the number of ligands increases the conformational penalty for tether stretching/compression near the cell surface and leads to a decrease in targeting efficiency. At the same time, using longer tethers or a larger core size allows ligands to interact with receptors over a larger cell surface area, which can enhance the nanoparticle affinity toward the cell surface. We also discuss the selectivity of nanoparticle targeting of cells with a high receptor density. Based on the obtained results, we provide recommendations for improving the nanoparticle binding affinity and selectivity, which can guide future nanoparticle development for diagnostic and therapeutic purposes.

Monte carlo simulations of metallo-supramolecular micelles.

Using Monte Carlo simulations we show that the equilibrium properties of metallo-supramolecular micelles are determined by the competition of 2:1 and 1:1 metal-ligand complexation in the bulk and on the surface as well as steric interactions between the neighboring corona blocks attached to the surface. We predict that by increasing the association energy for the second metal-ligand bond, or decreasing the corona block length one can achieve a larger core surface coverage for metallo-supramolecular micelles. Compared to covalently bonded block copolymer micelles, we show that metallo-supramolecular micelles have smaller monomer and end group density, especially in the vicinity of the core, which may lead to experimentally observed aggregation.

Cis-trans switchable metallosupramolecular polymers: computer modeling.

Using computer simulations we study linear oligomers end functionalized with ligands that can form trans- or cis-2:1 complexes with metal ions in a salt-screened good solvent. We show that trans-cis isomerization of ligand-metal complexes can significantly increase the average molecular weight as well as trigger formation of reversible metallosupramolecular network based on 3:1 ligand-metal complexes acting as cross-linkers. We predict the conditions under which the most dramatic changes in the properties of metallosupramolecular polymers, such as network formation or increase in elastic plateau modulus of the network, occur upon isomerization.

Optimization of functionalized polymer layers for specific targeting of mobile receptors on cell surfaces.

The reversible binding between a planar polymer layer functionalized by ligands and a planar cell surface containing different densities of mobile receptors has been studied by Monte Carlo simulations. Using the acceptance-ratio method, the distance-dependent profiles for the average number of ligands bound to receptors, the total free energy for the polymer layer-cell surface interaction and the interaction force were obtained. Four main design parameters for the polymer layer were considered: the degree of functionalization, chain degree of polymerization, polymer grafting density and the binding energy for the ligand-receptor interaction. We found that an increase in the degree of functionalization or in the absolute energy of ligand-receptor binding results in a larger number of ligands bound to the receptors, lower free energy, and stronger attractive force. Polymer layers composed of shorter chains were found to exhibit a deeper and narrower free energy profile and a larger attractive force, while longer tethers can interact with the cell surface at a larger and broader range of separation distances, in agreement with experimental observations. Our simulation results show that the increase in polymer grafting density from the mushroom to brush regime enhances the ligand availability and results in a stronger attractive force, increases the maximum binding distance, but exhibits a shallower free energy minimum due to the smaller tolerance to compression for polymer layers with high grafting density. We used two measures of the polymer layer binding affinity to the cell surface: the free energy minimum, related to the equilibrium binding constant and the fraction of bound ligands. We found that the polymer layers with a smaller chain length and grafting density, larger degree of functionalization, and larger absolute binding energy exhibit both a larger equilibrium binding constant to the cell surface and a larger average number of bound ligands, except for high binding energies when the maximum level of binding is reached independently of polymer length and grafting density. We showed that high binding specificity can be achieved by the polymer layers with intermediate ligand-receptor binding energies or an intermediate number of ligands, as a larger binding energy or number of ligands ensures a high binding affinity but lacks specificity while a smaller binding energy or number of ligands provides inadequate affinity. We found that the results for polymer layers with different properties follow a similar pattern when both high binding affinity to cells with high receptor density and high binding specificity are considered. As a result, the optimal design of the polymer layers can be achieved by using several different strategies, which are discussed.

Doxorubicin and beta-lapachone release and interaction with micellar core materials: experiment and modeling.

Polymer micelles with two different core-forming blocks, poly(d,l -lactide) (PLA) and poly(epsilon-caprolactone) (PCL), but the same coronal material, poly(ethylene glycol) (PEG), were investigated in this study as nanoscopic drug carriers. The release of two different drugs, doxorubicin (DOX) and beta-lapachone (beta-lap), from PEG(5k)-b-PCL(5k) and PEG(5k)-b-PLA(5k) micelles was studied at pH 5.0 and 7.4. Mathematical solutions of both Higuchi's model and Fickian diffusion equations were utilized to elucidate the differences between the micelle core materials for the two drugs. The neutral and smaller of the two drugs tested, beta-lap, demonstrated faster, pH-independent release, suggesting that no substantial changes occurred in either micelle core at lower pH. In contrast, the release rate of DOX was found to noticeably increase at lower pH with a larger cumulative amount of drug released. Different core materials were shown to have considerable influence on the release kinetics of both drugs: in both cases, the more hydrophobic PCL core showed slower drug release rates compared with the less hydrophobic PLA core.

[Liposuction in facial contour rejuvenation].

OBJECTIVE: To evaluate the effect of vacuum suction on facial shaping and cervicofacial rejuvenating. METHODS: Based on injection of swelling solution, vacuum suction was used in totally 327 cases, including local fat deposit in face or neck. RESULTS: Since 1993, we have carried out 327 case, satisfactory rate was up to 98%. CONCLUSIONS: This is an adaptable method for facial shaping and cervicofacial Rejuvenating, very good results can be achieved.

[Structural characteristics of infrared spectra for polyaluminum chloride in enhanced reactions of modification].

As an important preparation method of Inorganic Polymer Flocculants (IPFs) with all-round performance, the use of enhanced reactions of modification for Polyaluminum Chlorides (PAC) is a trend of application. In this paper, auctorial-made multicore oligomers as a complexing agent, polysilicate as a crosslinking agent and polyacrylamide as a compounding agent are respectively used to enhance PAC and prepared three kinds of IPFs, named CO-PAC, CR-PAC and PAM-PAC. The Fourier Transform Infrared spectra (FTIR) of the three flocculants show that there have been found new chemical bonds beyond Al-OH in their structures, and each characteristic band of primary group in PAC has a band shift to a lower frequency due to its structural aberration. These evidences suggested that enhanced reactions of modification for PAC are not at all a simple process of physical mix, but a chemical reaction to form other structural characteristics of inorganic polymers. The excellent coagulation behavior of these new flocculants depends mainly on the chemical actions, but there are a little physical effects between the modifiers and PAC.