Combining training data from multiple sources increases sample size and reduces confounding, leading to more accurate and less biased machine learning models. In healthcare, however, direct pooling of data is often not allowed by data custodians who are accountable for minimizing the exposure of sensitive information. Federated learning offers a promising solution to this problem by training a model in a decentralized manner thus reducing the risks of data leak-age. Although there is increasing utilization of federated learning on clinical data, its efficacy on individual-level genomic data has not been studied. This study lays the groundwork for the adoption of federated learning for genomic data by investigating its applicability in two scenarios: phenotype prediction on the UK Biobank data and ancestry prediction on the 1000 Genomes Project data. We show that federated models trained on data split into independent nodes achieve performance close to centralized models, even in the presence of significant inter-node heterogeneity. Additionally, we investigate how federated model accuracy is affected by communication frequency and suggest approaches to reduce computational complexity or communication costs.

We investigate the processes of microrelief formation on a Si(100) surface under irradiation with a Ga+-ion beam with an energy of 30 keV and a fluence of D = 1.25 × 1018–2 × 1019 cm–2 at incident angles of θ = 30°–85°. Within the angular range of θ = 40°–70°, a faceted wavy relief forms on the Si surface, while at θ = 30°, a sinusoidal relief develops. An experimental dependence of the periodic structure wavelength as a function of irradiation time λ(t) t n, where n = 0.33–0.35, is obtained. The average values of relief propagation velocities and their direction relative to the incident ion direction are determined for θ = 30° and 40°, amounting to –5.3 ± 0.6 and –6.3 ± 0.6 nm s–1, respectively. The results are discussed in detail within the framework of existing models of wavelike surface relief formation under ion bombardment.

The design of a laboratory solid-state thermostat based on Peltier elements with an operating temperature range from −50 to +90°C is described. The thermostat is made from up-to-date circuitry components. Thin films, silicon chip fragments with analyzed structures, and other miniature objects with lateral dimensions of up to 15 × 15 mm can be used as a test sample. The sample thickness is limited by the thermal conductivity of its material, and it can be as large as 3 mm.

We consider the equation $u_{xx}-u+W(x)u^3=0$ where $W(x)$ is a periodic alternating piecewise constant function. It is proved that under certain conditions for $W(x)$ solutions of this equation, which are bounded on $\mathbb{R}$, $|u(x)|<\xi$, can be put in one-to-one correspondence with bi-infinite sequences of numbers $n\in \{-N,\,\ldots,\,N\}$ (called ``codes'' of the solutions). The number $N$ depends on the bounding constant $\xi$ and the characteristics of the function $W(x)$. The proof makes use of the fact that, if $W(x)$ changes sign, then a ``great part'' of the solutions are singular, i.e., they tend to infinity at a finite point of the real axis. The nonsingular solutions correspond to a fractal set of initial data for the Cauchy problem in the plane $(u,\,u_x)$. They can be described in terms of symbolic dynamics conjugated with the map-over-period (monodromy operator) for this equation. Finally, we describe an algorithm that allows one to sketch plots of solutions by its codes.

Classifying the degree of relatedness between pairs of individuals has both scientific and commercial applications. As an example, genome-wide association studies (GWAS) may suffer from high rates of false positive results due to unrecognized population structure. This problem becomes especially relevant with recent increases in large-cohort studies. Accurate relationship classification is also required for genetic linkage analysis to identify disease-associated loci. Additionally, DNA relatives matching service is one of the leading drivers for the direct-to-consumer genetic testing market. Despite the availability of scientific and research information on the methods for determining kinship and the accessibility of relevant tools, the assembly of the pipeline, which stably operates on a real-world genotypic data, requires significant research and development resources. Currently, there is no open source end-to-end solution for relatedness detection in genomic data, that is fast, reliable and accurate for both close and distant degrees of kinship, combines all the necessary processing steps to work on a real data, and is ready for production integration. To address this, we developed GRAPE: Genomic RelAtedness detection PipelinE. It combines data preprocessing, identity-by-descent (IBD) segments detection, and accurate relationship estimation. The project uses software development best practices, as well as Global Alliance for Genomics and Health (GA4GH) standards and tools. Pipeline efficiency is demonstrated on both simulated and real-world datasets. GRAPE is available from: https://github.com/genxnetwork/grape.

Experimental studies of the germanium surface morphology development under irradiation with a focused gallium ion beam at different angles of incidence and fluences are presented. It is shown that a nanoporous structure formes in the near-surface layer starting with a dose of 5.10(15) cm(-2).This leads to the formation of a sponge- like morphology with a wall thickness of about 20 nm and a depth up to 150 nm with an increasing dose. Changing the ion beam incidence angle with respect to the surface normal leads to a tilt of the pores walls in the collinear direction.

Classifying the degree of relatedness between pairs of individuals has both scientific and commercial applications. As an example, genome-wide association studies (GWAS) may suffer from high rates of false positive results due to unrecognized population structure. This problem becomes especially relevant with recent increases in large-cohort studies. Accurate relationship classification is also required for genetic linkage analysis to identify disease-associated loci. Additionally, DNA relatives matching service is one of the leading drivers for the direct-to-consumer genetic testing market. Despite the availability of scientific and research information on the methods for determining kinship and the accessibility of relevant tools, the assembly of the pipeline, which stably operates on a real-world genotypic data, requires significant research and development resources. Currently, there is no open source end-to-end solution for relatedness detection in genomic data, that is fast, reliable and accurate for both close and distant degrees of kinship, combines all the necessary processing steps to work on a real data, and is ready for production integration. To address this, we developed GRAPE: Genomic RelAtedness detection PipelinE. It combines data preprocessing, identity-by-descent (IBD) segments detection, and accurate relationship estimation. The project uses software development best practices, as well as Global Alliance for Genomics and Health (GA4GH) standards and tools. Pipeline efficiency is demonstrated on both simulated and real-world datasets. GRAPE is available from: https://github.com/genxnetwork/grape.

The silicon surface was irradiated with 30 keV gallium ion beam at incidence angles from 0 to 50 and fluences from 6middot10(16) to 5middot10(18) cm(-2). Surface topography was investigated by scanning electron microscopy. It was found that one of four types of a relief can be formed on the silicon surface depending on the ion beam incidence angle and these fluences. Pattern formation starts with fluences of ~2middot10(17 )cm(-2). The peculiarities of a relief evolution can be explained by the angular dependences of silicon sputtering with gallium ion beam and the possible existence of implanted gallium in the near-surface layer in the form of precipitates.

The results of investigation of the polarization relaxation mechanism for the LiPON solid electrolyte by a discharge through an external load are presented. Test cells implemented in the form of encapsulated multilayer structures $${{{{{{\text{Si}{{\text{O}}_{\text{2}}}}/{\text{Pt}}(100\,\,\text{nm})}/{\text{LiPON}\,(1000\,\,\text{nm})}}/{\text{Pt}~(100\,\,\text{nm})}}/{\text{Ti}~(10\,\,\text{nm})}}/{\text{Si}{{\text{O}}_{\text{2}}}}}/{\text{Si}}\;$$ and previously studied by the standard techniques are tested on a special bench measuring discharge characteristics. The core of the method lies in charging the test cell from a stable voltage source up to saturation with the subsequent rapid switch to precision resistance and detection of the voltage drop. The measurements are taken in the load range from 0.1 MΩ to 10 Ω in the temperature range of –50 to 25°C. An equivalent electric circuit for the test cell is proposed, for which a mathematical model of the discharge process is constructed. In the context of the proposed model, the features of the experimental curves are explained by the processes of redox reactions of the lithium ions at the surface of the electrodes and generation of nonequilibrium charge carriers in the bulk upon the charging and discharging of the test cell.

We study localized modes (LMs) of the one-dimensional Gross-Pitaevskii/nonlinear Schrödinger equation with a harmonic-oscillator (parabolic) confining potential, and a periodically modulated coefficient in front of the cubic term (nonlinear lattice pseudopotential). The equation applies to a cigar-shaped Bose-Einstein condensate loaded in the combination of a magnetic trap and an optical lattice which induces the periodic pseudopotential via the Feshbach resonance. Families of stable LMs in the model feature specific properties which result from the interplay between spatial scales introduced by the parabolic trap and the period of the nonlinear pseudopotential. Asymptotic results on the shapes and stability of LMs are obtained for small-amplitude solutions and in the limit of a rapidly oscillating nonlinear pseudopotential. We show that the presence of the lattice pseudopotential may result in: (i) creation of new LM families which have no counterparts in the case of the uniform nonlinearity; (ii) stabilization of some previously unstable LM species; (iii) evolution of unstable LMs into a pulsating mode trapped in one well of the lattice pseudopotential.

We consider finite temperature effects in a non-standard Bose-Hubbard model for an exciton- polariton Josephson junction (JJ) that is characterised by complicated potential energy landscapes (PEL) consisting of sets of barriers and wells. We show that the transition between thermal activation (classical) and tunneling (quantum) regimes exhibits universal features of the first and second order phase transition (PT) depending on the PEL for two polariton condensates that might be described as transition from the thermal to the quantum annealing regime. In the presence of dissipation the relative phase of two condensates exhibits non-equilibrium PT from the quantum regime characterized by efficient tunneling of polaritons to the regime of permanent Josephson or Rabi oscillations, where the tunneling is suppressed, respectively. This analysis paves the way for the application of coupled polariton condensates for the realisation of a quantum annealing algorithm in presently experimentally accessible semiconductor microcavities possessing high (10 5 and more) Q-factors.

Представлены результаты исследования тонкопленочных образцов твердого электролита LiPON, полученных методом магнетронного нанесения. Приведены данные о морфологии, структуре, элементном и фазовом составе пленок LiPON и результаты исследования их электрофизических параметров. DOI: 10.21883/PJTF.2017.11.44690.16637

The results of developing the technology of magnetron sputtering deposition of a LiPON solid electrolyte and an experimental investigation of its characteristics are presented. The basic processing operations and parameters providing the formation of films of the proper morphology, structure, and elemental and phase composition are described. The data of the measurement of the physical parameters of the films by cyclic voltammetry and potentiometry are represented.

Results of studies of the solid electrolyte effect on capacitance of thin-film electrodes on the basis of Si-O-Al and VxOy nanocomposites are presented. The studies were carried out by comparing the charge-discharge characteristics of two pairs of the identical electrodes, one of which was covered by LiPON film, within prototypes with two lithium electrodes - the counter and the reference electrode.

The results of examination of thin-film samples of the LiPON solid electrolyte, which were synthesized by magnetron sputtering, are reported. Data on the morphology, structure, elemental and phase composition, and electrophysical parameters of LiPON films are presented.