Limited Quantity Of Data Research Articles

PCSboost: A Multi-Model Machine Learning Framework for Key Fragments Selection of Channelrhodopsins Achieving Optogenetics

Optogenetics combines optical and genetic methods to modulate light-controlled gene expression, protein localization, signal transduction and protein interactions to achieve precise control of specific neuronal activity, with the advantages of low tissue damage, high spatial and temporal resolution, and genetic specificity. It provides a cutting-edge approach to establishing a causal relationship between brain activity and behaviors associated with health and disease. Channelrhodopsin (ChR) functions as a photogenic activator for the control of neurons. As a result, ChR and its variants are more widely used in the realization of optogenetics. To enable effective optogenetics, we propose a novel multi-model machine learning framework, i.e., PCSboost, to accurately assist key fragments selection of ChRs segments that realize optogenetics from protein sequence structure and information dataset. We investigate the key regions of the ChR variant protein fragments that impact photocurrent properties of interest and automatically screen important fragments that realize optogenetics. To address the issue of the dataset containing a limited quantity of data but a high feature dimension, we employ principal component analysis (PCA) to reduce the dimensionality of the data and perform feature extraction, followed by the XGBoost model to classify the ChRs based on their kinetics, photocurrent and spectral properties. Simultaneously, we employ the SHAP interpretability analysis to perform an interpretability analysis of the ChR variant protein for pointwise, characteristic similarities to identify key regions of the protein fragment structure that contribute to the regulation of photocurrent intensity, photocurrent wavelength sensitivity and nonkinetic properties. Experimental findings demonstrate that our proposed PCSboost approach can speed up genetic and protein engineering investigations, simplify the screening of important protein fragment sections, and potentially be used to advance research in the areas of optogenetics, genetic engineering and protein engineering.

Automatic assessment of mammographic density using a deep transfer learning method.

Mammographic breast density is one of the strongest risk factors for cancer. Density assessed by radiologists using visual analogue scales has been shown to provide better risk predictions than other methods. Our purpose is to build automated models using deep learning and train on radiologist scores to make accurate and consistent predictions. We used a dataset of almost 160,000 mammograms, each with two independent density scores made by expert medical practitioners. We used two pretrained deep networks and adapted them to produce feature vectors, which were then used for both linear and nonlinear regression to make density predictions. We also simulated an "optimal method," which allowed us to compare the quality of our results with a simulated upper bound on performance. Our deep learning method produced estimates with a root mean squared error (RMSE) of . The model estimates of cancer risk perform at a similar level to human experts, within uncertainty bounds. We made comparisons between different model variants and demonstrated the high level of consistency of the model predictions. Our modeled "optimal method" produced image predictions with a RMSE of between 7.98 and 8.90 for cranial caudal images. We demonstrated a deep learning framework based upon a transfer learning approach to make density estimates based on radiologists' visual scores. Our approach requires modest computational resources and has the potential to be trained with limited quantities of data.

Fractal Informed Generative Adversarial Networks (FI-GAN): Application to the generation of X-ray CT images of a self-similar partially saturated sand

The recent increase in data availability fostered the development of new and more powerful data-driven numerical models, particularly machine learning. With a limited quantity of data in geomechanics applications, machine learning often suffers from unstable training and lack of convergence. Physics informed machine learning models have been proposed in the literature to mitigate the issue as an alternative to data-only approaches. In the present work, a new machine learning method is proposed: Fractal Informed Generative Adversarial Networks (FI-GAN) and applied to a case of X-ray CT images of partially saturated sand. This method consists of training an informed adversarial model using pore fractal dimension. The semi-supervision provided by pore fractal metrics presents two advantages, firstly, it regularises the training; secondly, it improves the distribution of pore fractal dimension, which is correlated with permeability. Results show that the generated images improve the fractal distribution of the informer phase in the partially saturated material (i.e., water) and the other two phases. The approach is validated with physics flow simulations. The generated images can be applied in the calibration of image processing tools, filling missing data in X-ray CT scans, generating microscales in multiscale applications, and data augmentation, among others.

Close-proximity, conservative extrapolation of load spectra

PurposeThe purpose of this paper is to present the author’s method of conservative load spectrum (LS) derivation and close-proximity LS extrapolation applying a correction for measurement uncertainty caused by too low sampling frequency or signal noise, which may affect the load histories collected during the flying session and cause some recorded load increments to be lower than the actual values.Design/methodology/approachHaving in mind that the recorded load signal is burdened with some measurement error, a conservative approach was applied during qualification of the recorded values into 32 discrete load-level intervals and derivation of 32 × 32 half-cycle arrays. A part of each cell value of the half-cycle array was dispersed into the neighboring cells placed above by using a random number generator. It resulted in an increase in the number of load increments, which were one or two intervals higher than those resulting from direct data processing. Such an array was termed a conservative clone of the actual LS. The close-proximity approximation consisted of multiplication of the LSs clones and their aggregation. This way, the LS for extended time of operation was obtained. The whole process was conducted in the MS Excel environment.FindingsFatigue life calculated for a chosen element of aircraft structure using conservative LS is about 20%–60% lower than for the actual LS (depending on the applied value of dispersion coefficients used in the procedure of LSs clones generation). It means that such a result gives a bigger safety margin when operational life of the aircraft is estimated or when the fatigue test for an extended operational period is programed based on a limited quantity of data from a flying session.Originality/valueThis paper presents a proposal for a novel, conservative approach to fatigue life estimation based on the short-term LS derived from the load signal recorded during the flying session.

Home fortification of foods with multiple micronutrient powders for health and nutrition in children under two years of age.

Home fortification of foods with multiple micronutrient powders for health and nutrition in children under two years of age.

Transfer Learning to Model Inertial Confinement Fusion Experiments

Inertial confinement fusion (ICF) experiments are designed using computer simulations that are approximations of reality and therefore must be calibrated to accurately predict experimental observations. In this article, we propose a novel technique for calibrating from simulations to experiments, or from low fidelity simulations to high fidelity simulations, via “transfer learning” (TL). TL is a commonly used technique in the machine learning community, in which models trained on one task are partially retrained to solve a separate, but related task, for which there is a limited quantity of data. We introduce the idea of hierarchical TL, in which neural networks trained on low fidelity models are calibrated to high fidelity models, then to experimental data. This technique essentially bootstraps the calibration process, enabling the creation of models which predict high fidelity simulations or experiments with minimal computational cost. We apply this technique to a database of ICF simulations and experiments carried out at the Omega laser facility. TL with deep neural networks enables the creation of models that are more predictive of Omega experiments than simulations alone. The calibrated models accurately predict future Omega experiments, and are used to search for new, optimal implosion designs.

K-NN classification of handwritten characters via accelerated GAT correlation

This paper addresses the problem of reinforcing the ability of the k-NN classification of handwritten characters via distortion-tolerant template matching techniques with a limited quantity of data. We compare three kinds of matching techniques: the conventional simple correlation, the tangent distance, and the global affine transformation (GAT) correlation. Although the k-NN classification method is straightforward and powerful, it consumes a lot of time. Therefore, to reduce the computational cost of matching in k-NN classification, we propose accelerating the GAT correlation method by reformulating its computational model and adopting efficient lookup tables. Recognition experiments performed on the IPTP CDROM1B handwritten numerical database show that the matching techniques of the simple correlation, the tangent distance, and the accelerated GAT correlation achieved recognition rates of 97.07%, 97.50%, and 98.70%, respectively. The computation time ratios of the tangent distance and the accelerated GAT correlation to the simple correlation are 26.3 and 36.5 to 1.0, respectively.

Candidate mechanisms underlying effects of contractile activity on muscle morphology and energetics in cancer cachexia

Skeletal muscle wasting is a prominent pathophysiological feature of cancer cachexia. Recent evidence suggests the manifestation of mitochondrial dysfunction along with a diminished oxidative capacity. These abnormalities have been concurrently observed with impaired muscle force production and the accelerated onset of fatigue in both tumour-bearing animals and cancer patients exhibiting wasting. To address the burden imposed by cachexia, nutritional and pharmacological interventions have been investigated extensively; in contrast, contractile activity-based countermeasures (i.e. exercise training) have been less frequently explored. Although limited, several preclinical studies that implemented contractile activity have reported favourable outcomes such as the retention of muscle mass and the restoration of energetic homeostasis. Even fewer investigations have examined the mechanisms accounting for these protective effects. An experimental approach addressing contractile activity-dependent expression of muscle mass and energy metabolism regulators may yield information that provides mechanistic support for exercise countermeasures. In this review, we present several candidate mechanisms underlying the protective effects of contractile activity as support for exercise countermeasure strategies. Given the limited quantity of data in this area, insights will be derived from studies on contractile activity-dependent modulation of common cellular and molecular events regulating muscle morphology and energetics during other muscle wasting conditions (e.g. sarcopenia).

Crustal structure variation in western Turkey inferred from the receiver function analysis

We have used receiver function analysis employing teleseismic waveforms to investigate variations in crustal and upper-mantle velocity structure beneath the western Turkey. The genetic algorithm is adopted to avoid the inherent non-uniqueness problems of the inversion. The total of 50 teleseismic waveforms recorded by 25 permanent stations is analyzed and the selected records of earthquakes distributed in three quadrants from each seismic station in the years from 2007 to 2008 are analyzed. 2-D images of Moho variation beneath the western Turkey along two profiles are 700 and 300 km long and extended from the Aegean Sea coast to the central Anatolia and southeastern Anatolia respectively are estimated. The depth-migrated receiver function showed almost well resolved Moho for the two profiles. Although the limited quantity of data is used to short periods, the resolution and confidence level is not so bad. For the one profile that extended from the Aegean Sea coast to central Anatolia, we obtain Moho depths between 25 and 35 km. On the other profile along the western coast, the Moho depths range from 20 and 30 km. Shear wave speed was found to vary between 3.3 km/s and 3.7 km/s in the lower crust, and between 4.1 km/s and 4.4 km/s uppermost mantle. We concluded that the Moho depth is increasing from western Anatolia to central Anatolia considering inversion results and 2-D depth-migrated receiver functions.

1D velocity structure beneath broadband seismic stations in the Cretaceous Gyeongsang Basin of Korea by receiver function analyses

The crustal velocity structures beneath four broadband seismic stations (GKP, GSU, HDB, and BUS) in the Cretaceous Gyeongsang Basin, southeastern Korea, are estimated by using receiver function analyses employing teleseismic waveforms. The genetic algorithm is adopted to avoid the inherent non-uniqueness problem of the inversion. The inversion results are constrained by surface-wave dispersions to complement the shortcoming of the receiver function. The selected records of earthquakes distributed in three quadrants from each seismic station in the years from 2002 to 2005 are analyzed. Although limited quantity of data is used due to short operation periods, the resolution and confidence level are expected to be better than those using less data. The Moho depth under GKP is estimated to be 30 km showing no distinctive indication of dipping. This depth is well coincident with those in nearby velocity cross-sections obtained from crustal-scale seismic profiles compared to the 36-km deep Moho previously reported by other independent work. At GSU the Moho appears at the 32-km depth and a distinct low-velocity anomaly is detected at the depth of about 10 km. They agree with those in the velocity tomogram of the nearby survey line. An adakitic intrusion which results from the partial melting of a young and hot subducted oceanic crust in the basin during the Cretaceous is suggested as a possible geologic interpretation of the low-velocity zone. The Moho beneath HDB is 28-km deep, and agrees with those in nearby velocity cross-sections obtained from crustal-scale seismic profiles. This Moho depth at HDB is rather shallower than those of other stations. The significant P S phase amplitude and arrival time differences in the radial receiver functions conclude the Moho to dip southwestwardly. This is supported by polarities of direct P and P S in transverse receiver functions. Beneath BUS, P wave velocity increases gradually from the 32-km depth and it reaches 7.6 km/s at the 36-km depth. The Moho discontinuity beneath BUS is thought to be at an around 35-km depth.

Assessment of Oil Spill Risk for Shuttle Tankers in the Gulf of Mexico

To date, all deepwater oil reserves in the Gulf of Mexico have been developed with systems that transport oil to shore via pipelines. This paper describes the methodology used to assess the oil spili risks for a feasible alternative to these systems, a tanker-based Floating Production Storage and Offloading (FPSO) system with oil transportation to shore via shuttle tankers. This methodology provided a rational and consistent framework for combining available historical data with input from technical experts. The methodology tracked all potential sources of uncertainty, including uncertainties due to a limited quantity of data (particularly for large spill incidents), uncertainties due to any adjustments made to the data to reflect current operating practices, and uncertainties due to extrapolating data reflecting current operating practices to predict future performance. The major conclusions are that the oil spill risk for shuttle tankers is dominated by rare, large spills rather than frequent, small spills; that the confidence intervals in the results are very wide, reflecting the limited quantity and quality of historical data available to estimate frequencies for rare events; and that the results from this study should be periodically updated because they provide a valuable baseline for future analyses of pollution risk in the Gulf of Mexico.

Étude des modèles thermodynamiques pour représenter des mélanges contenant des hydrocarbures, de l’eau et des alcools: II. Prédiction des propriétés thermodynamiques des systèmes binaires à partir d’une base restreinte de données expérimentales

In this work we have been interested in the problem of predicting thermodynamic properties of binary systems in a wide range of temperature or pressure. Towards this end, we have estimated the parameters of the models considered in a previous paper by using both vapor–liquid equilibrium data and mixing enthalpies determined at low temperatures (usually at 298.15 K). The obtained results show that at moderates temperatures, the representation of the vapor–liquid equilibria and mixing enthalpies is comparable to that obtained with the parameters estimated over the whole set of data used in the first part of this work. At high temperature, the calculation of the vapor–liquid equilibria is very satisfactory. The interest of this method remains, of course, in the limited quantity of data which we can dispose (i.e., thermodynamic properties at a single temperature), but also in the method of parameter estimation that ensure the thermodynamic consistency of the models.

New data on the crystal-chemistry of fluoborite by means of SREF, SIMS, and EMP analysis

The crystal structure of fluoborite [Mg 3 F 3 (BO 3 )] was refined by Dal Negro and Tadini (1974) who provided a complete structural model. Previously, Takeuchi (1950) had refined an OH-dominant fluoborite (OH ~70%), but the limited quantity of data (extracted from two Weissenberg-Buerger photographs) did not permit the location of H atoms. Dal Negro and Tadini (1974) also could not locate H atoms because they used a crystal with near end-member composition. We have located the H bond in an OH-dominant fluoborite from the Betic Cordilleras (SE Spain). Excellent quality X-ray data on two crystals of fluoborite allowed discovery and refinement of the H position in this mineral. Electron microprobe (EMP) and secondary-ion mass spectrometry (SIMS) analyses of the light elements H, B, and F have resulted in the formulation of special procedures to obtain accurate, high-quality quantitative data, which are presented in this paper. EMP, SIMS, and crystal structure refinement (SREF) data are in a good agreement. Linear equations are also presented to calculate the F content directly from cell parameters.

RESEARCH NOTE ESTIMATING SHEETING ROLL CLOSING FORCES THROUGH MEASUREMENT of ROLL POWER CONSUMPTION

An important factor in the design of the sheeting operation used in the snack food and bakery industries are the forces exerted on the rolls as a result of the pressure developed between the rolls. These forces can cause significant mechanical deflections that result in nonuniform products being produced across the width of the rolls. Unfortunately, measurements of closing forces on operating sheeting systems can be difficult to obtain. This note illustrates a simple method for estimating the difficult to obtain sheeter closing forces from more easily obtained power measurements. A limited quantity of data confirms the accuracy of the procedure.

Relating C:N ratios in Zooplankton food and faecal pellets using a biochemical model

A model is developed which predicts faecal pellet C:N ratio by considering how carbon and nitrogen in food items are distributed between various biochemical compounds (proteins, carbohydrates, lipids), and then how these compounds are assimilated within the Zooplankton gut. The model successfully reproduced the generally observed trend that pellet C:N ratios are greater than those of the food from which they are derived. The model is superior to simpler models based on elements rather than specific compounds because it has the potential to reproduce the wide variability in average carbon assimilation efficiency which occurs among foods of different types. The model was, however, derived from a very limited quantity of data. The work highlighted the need for more data on the biochemical composition of marine organisms, and how various biochemical fractions are assimilated by Zooplankton, in order to provide realistic predictions of how C:N ratios in pellets relate to those in food.

The copper content of foods based on a critical evaluation of published analytical data

Increasing interest in the copper (Cu) intake of Americans has generated a need for the compilation, evaluation, and improvement of data for Cu in foods. In order to estimate dietary Cu intake, accurate and precise Cu values for foods are needed. A system of criteria developed to evaluate the quality of published selenium data has been adapted to evaluate analytical data for Cu in foods. Mean Cu values for each food were calculated from the evaluated data and combined with USDA frequency of consumption data, resulting in a list of 218 major contributors of dietary Cu. The richest sources of Cu are legumes, wheat bran cereals, organ meats, shellfish, and grains. Confidence codes, indicators of the relative degree of confidence the user of the data can have in that mean value, were included. More than half of the mean values for the foods listed are of limited quality or result from a limited quantity of data, indicating a need for improvement in food Cu data. This proposed dynamic system for the compilation and evaluation of Cu data can be used to generate Cu data bases for specific purposes, provide a ranked list of foods which are significant contributors of that nutrient, and establish priorities for further improvements in the database.

Thermodynamics of binary and ternary ion exchange in zeolites: the exchange of sodium, ammonium and potassium ions in mordenite

In many ion exchanges that involve zeolites, more than two types of exchanging ions are involved. Most work to date has been concerned with equilibria involving two different types of ion only. In this paper, a recently devised thermodynamic treatment for exchange equilibria involving three ions is tested experimentally. In § 1 the most important aspects of this thermodynamic model are discussed, and after this the experimental techniques necessary to obtain a ternary exchange isotherm are summarized. Data for three binary exchanges in synthetic mordenite are presented in §3 (the ion pairs Na-NH 4 , NA-K and K-NH 4 ) together with a ternary isotherm obtained for the simultaneous exchange of Na+, NH +4 and K + ions in the same zeolite, all data being measured at a temperature of 25 °C. After some general considerations in §4 on the nature of non-ideality in the zeolite phase and its effects on the variation of observed ion selectivities as a function of exchanger composition, the experimental data are next used to test, first, the validity, and secondly the range of applicability of numerous models found in the literature, all of which attempt to predict multicomponent exchange equilibria from binary exchange data alone. In §5 models that were developed primarily with ion exchange resins in mind are discussed (Soldatov & Bychkova 1980). In §6 approaches that attempt to predict, from binary activity data, the coefficients of ternary exchanger systems are considered, especially those that have been used with some success on clays (Elprince & Babcock 1975; Chu & Sposito 1981). All of these models are shown to be, at best, only partially successful in predicting experimental data for the comparatively simple Na-NH 4 -K-mordenite exchange system. The use of zeolites seems, therefore, to introduce additional complications. The recently developed model of Fletcher & Townsend (1981 b is next used to determine directly the exchanger phase activity coefficients 0 Na, 0 NH 4 and 0 K , and these are indeed found to vary with the exchanger phase composition in a complicated manner. Finally, in §8, the observed non-ideality of the ternary exchange system is rationalized in terms of site heterogeneity within the zeolite framework, and the consequences of this phenomenon are discussed, particularly with respect to the problems this presents in the successful prediction of multicomponent exchange phenomena, when only a limited quantity of data are available.

Lepton spectrum in the semileptonic weak decays of charmed mesons

If we assume that the semileptonic decays of the charmed $D$ mesons are dominated by the two channels $D\ensuremath{\rightarrow}\mathrm{Kl}\ensuremath{\nu}$ and $D\ensuremath{\rightarrow}{K}^{*}l\ensuremath{\nu}$, then, by measuring the average energy of the decay lepton, we can put an upper bound on the ratio $R=\frac{\ensuremath{\Gamma}(D\ensuremath{\rightarrow}\mathrm{Kl}\ensuremath{\nu})}{\ensuremath{\Gamma}(D\ensuremath{\rightarrow}{K}^{*}l\ensuremath{\nu})}$. Using a limited quantity of data from DESY, we find $R\ensuremath{\lesssim}0.7$.

Measurements of extreme concentrations of tropospheric hydrogen sulfide

The ambient concentration of hydrogen sulfide, a species of significance to both gas phase tropospheric chemistry and the interior chemistry of atmospheric aerosols, has been measured over a 13-month period in a location representative of an extreme H2S environment. Peak concentrations exceeding 1 ppm (20-min average) and long-term average values of 24 and 12 ppb (for two different source configurations) were measured at the Artesia Junction, New Mexico, site: such levels exceed nearly all known ambient values by factors of 10–20. The concentration data are found to be represented (in different cases) by either Weibull or gamma distributions or by beta or log normal distributions, this finding thus contradicting previous suggestions of universal log normality in air pollution data and rendering inappropriate the attempts to predict maximum concentrations by the imposition of unverified concentration distributions on limited quantities of data. Nearly all extremely high concentrations occurred between 1800 and 600 hours and were strongly associated with minimums in the local wind turbulence spectrum. The ‘contaminant exposure diagrams’ and the concentration distributions that are presented constitute the first detailed determinations of ambient hydrogen sulfide concentrations.

Fast time resolved spectral analysis of VLF banded emissions

Several classes of VLF banded emissions acquired by the short antenna Ogo 5 electric field detector are investigated by using high time resolution spectral techniques. The fastest temporal resolution of statistically independent spectra is 50 msec; individual spectra are taken every 12.5 msec. VLF banded rising chorus spectra possessed single and/or multiple frequency modulated bursts; discrete single modes with unchanging center frequencies are often observed to be part of the banded chorus spectrum. Inverted or falling banded chorus spectra also contained frequency modulated bursts and single modes, but with greater structural complexity than rising chorus. The spectrum of a banded ‘hisslike’ emission was dominated by multiple discrete single modes with constant center frequencies. A narrowband emission with frequencies slightly above one-half the local electron cyclotron frequency consisted of single modes and groups of unresolved modes that were not frequency modulated. Finally an emission at 3/2 the local electron cyclotron frequency exhibited an almost uniform spectral density across its 200-Hz bandwidth. Since this preliminary study contains only a limited quantity of data, the reported spectral structures probably neither exhaust nor are necessarily typical of all classes of banded VLF emissions.