Optimizing Nondecomposable Data Dependent Regularizers via Lagrangian Reparameterization Offers Significant Performance and Efficiency Gains.
Ontology highlight
ABSTRACT:
SUBMITTER: Ravi SN
PROVIDER: S-EPMC8174799 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Functional magnetic resonance imaging (fMRI) is a powerful, noninvasive tool that has significantly contributed to the understanding of the human brain. FMRI data provide a sequence of whole-brain volumes over time and hence are inherently four dimensional (4D). Missing data in fMRI experiments arise from image acquisition limits, susceptibility and motion artifacts or during confounding noise removal. Hence, significant brain regions may be excluded from the data, which can seriously undermine the quality of subsequent analyses due to the significant number of missing voxels. We take advantage of the four dimensional (4D) nature of fMRI data through a tensor representation and introduce an effective algorithm to estimate missing samples in fMRI data. The proposed Riemannian nonlinear spectral conjugate gradient (RSCG) optimization method uses tensor train (TT) decomposition, which enables compact representations and provides efficient linear algebra operations. Exploiting the Riemannian structure boosts algorithm performance significantly, as evidenced by the comparison of RSCG-TT with state-of-the-art stochastic gradient methods, which are developed in the Euclidean space. We thus provide an effective method for estimating missing brain voxels and, more importantly, clearly show that taking the full 4D structure of fMRI data into account provides important gains when compared with three-dimensional (3D) and the most commonly used two-dimensional (2D) representations of fMRI data.
| S-EPMC8612463 | biostudies-literature
Project description:The current commonly used single-guide RNA (sgRNA) structure has a shortened duplex compared with the native bacterial clustered regularly interspaced short palindromic repeats RNA (crRNA)–transactivating crRNA (tracrRNA) duplex. Here we show that modifying the sgRNA structure by extending the duplex length and mutating the fourth T of the continuous sequence of Ts (which is the pause signal for RNA polymerase III [pol III]) to C or G significantly, and sometimes dramatically, improves knockout efficiency in cells. In addition, the new sgRNA structure also significantly increases the efficiency of more challenging genome-editing procedures, such as gene deletion, which is important for inducing a loss-of-function in non-coding genes.
2015-11-17 | GSE74766 | GEO
Project description:ObjectiveUsing the American College of Rheumatology Rheumatology Informatics System for Effectiveness (RISE) registry, our objective was to examine performance on rheumatoid arthritis (RA) quality measures and to assess the association between practice characteristics and changes in performance over time among participating practices.MethodsWe analyzed data from practices enrolled in RISE between January 1, 2015 and December 31, 2017. Eight quality measures in the areas of RA disease management, cardiovascular risk reduction, and patient safety were examined. Variability in performance was evaluated at the practice level. Multivariate linear models were used to predict change in measure performance by year and to determine the effect of practice characteristics on change in performance over time.ResultsData from 59,986 patients from 54 practices were examined. The mean ± SD age was 62 ± 14 years, 77% were female, 69% were Caucasian, and most patients were seen in a single-specialty group practice (46%). The average performance on measures related to RA treatments was consistently high (>90%) across the study period. Measures related to RA functional status and disease activity assessment had the greatest improvements over time (8.4% and 13.0% increase per year, respectively; P < 0.001). Single-specialty group practices had the fastest rates of improvement over time across all measures.ConclusionAmong practices participating in RISE between 2015 and 2017, performance on most RA quality measures improved. Single-specialty group practices saw the fastest rates of improvement over time. Identification of workflow patterns leading to dramatic improvements in quality of care will help guide process redesign to address gaps in priority areas, such as tuberculosis screening and blood pressure control.
| S-EPMC7960552 | biostudies-literature
Project description:Breeding programs aiming to improve the performance of crossbreds may benefit from genomic prediction of crossbred (CB) performance for purebred (PB) selection candidates. In this review, we compared genomic prediction strategies that differed in 1) the genomic prediction model used or 2) the data used in the reference population. We found 27 unique studies, two of which used deterministic simulation, 11 used stochastic simulation, and 14 real data. Differences in accuracy and response to selection between strategies depended on i) the value of the purebred crossbred genetic correlation (rpc), ii) the genetic distance between the parental lines, iii) the size of PB and CB reference populations, and iv) the relatedness of these reference populations to the selection candidates. In studies where a PB reference population was used, the use of a dominance model yielded accuracies that were equal to or higher than those of additive models. When rpc was lower than ~0.8, and was caused mainly by G × E, it was beneficial to create a reference population of PB animals that are tested in a CB environment. In general, the benefit of collecting CB information increased with decreasing rpc. For a given rpc, the benefit of collecting CB information increased with increasing size of the reference populations. Collecting CB information was not beneficial when rpc was higher than ~0.9, especially when the reference populations were small. Collecting only phenotypes of CB animals may slightly improve accuracy and response to selection, but requires that the pedigree is known. It is, therefore, advisable to genotype these CB animals as well. Finally, considering the breed-origin of alleles allows for modeling breed-specific effects in the CB, but this did not always lead to higher accuracies. Our review shows that the differences in accuracy and response to selection between strategies depend on several factors. One of the most important factors is rpc, and we, therefore, recommend to obtain accurate estimates of rpc of all breeding goal traits. Furthermore, knowledge about the importance of components of rpc (i.e., dominance, epistasis, and G × E) can help breeders to decide which model to use, and whether to collect data on animals in a CB environment. Future research should focus on the development of a tool that predicts accuracy and response to selection from scenario specific parameters.
| S-EPMC8499581 | biostudies-literature
Project description:Nitrogen is essential for plant growth, and N fertilization allows farmers to obtain high yields and produce sufficient agricultural commodities. On the other hand, nitrogen losses potentially cause adverse effects to ecosystems and to human health. Increasing nitrogen use efficiency (NUE) is vital to solve the conflict between productivity, to secure the demand of a growing world population, and the protection of the environment. To ensure this, genetic improvement is considered to be a paramount aspect toward ecofriendly crop production. Winter oilseed rape (Brassica napus L.) is the second most important oilseed crop in the world and is cultivated in many regions across the temperate zones. To our knowledge, this study reports the most comprehensive field-based data generated to date for an empirical evaluation of genetic improvement in winter oilseed rape varieties under two divergent nitrogen fertilization levels (NFLs). A collection of 30 elite varieties registered between 1989 and 2014, including hybrids and open pollinated varieties, was tested in a 2-year experiment in 10 environments across Germany for changes in seed yield and seed quality traits. Furthermore, NUE was calculated. We observed a highly significant genetics-driven increase in seed yield per-se and, thus, increased NUE at both NFLs. On average, seed yield from modern open-pollinated varieties and modern hybrids was higher than from old open-pollinated varieties and old hybrids. The annual yield progress across all tested varieties was ~35 kg ha-1 year-1 at low nitrogen and 45 kg ha-1 year-1 under high nitrogen fertilization. Furthermore, in modern varieties an increased oil concentration and decreased protein concentration was observed. Despite, the significant effects of nitrogen fertilization, a surprisingly low average seed yield gap of 180 kg N ha-1 was noted between high and low nitrogen fertilization. Due to contrary effects of N fertilization on seed yield per-se and seed oil concentration an oil yield of 2.04 t ha-1 was measured at both N levels. Collectively, the data reveal that genetic improvement through modern breeding techniques in conjunction with reduced N fertilizer inputs has a tremendous potential to increase NUE of oilseed rape.
| S-EPMC5461335 | biostudies-literature
Project description:Evanescent wave absorption-based mid-infrared chalcogenide fiber sensors have prominent advantages in multicomponent liquid and gas detection. In this work, a new approach of tapered-fiber geometry optimization was proposed, and the evanescent efficiency was also theoretically calculated to evaluate sensing performance. The influence of fiber geometry (waist radius (Rw), taper length (Lt), waist deformation) on the mode distribution, light transmittance (T), evanescent proportion (TO) and evanescent efficiency (τ) is discussed. Remarkably, the calculated results show that the evanescent efficiency can be over 10% via optimizing the waist radius and taper length. Generally, a better sensing performance based on tapered fiber can be achieved if the proportion of the LP11-like mode becomes higher or Rw becomes smaller. Furthermore, the radius of the waist boundary (RL) was introduced to analyze the waist deformation. Mode proportion is almost unchanged as the RL increases, while τ is halved. In addition, the larger the micro taper is, the easier the taper process is. Herein, a longer waist can be obtained, resulting in larger sensing area which increases sensitivity greatly.
| S-EPMC9181228 | biostudies-literature
Project description:BackgroundComplementary-DNA based amplified fragment length polymorphism (cDNA-AFLP) is a commonly used tool for assessing the genetic regulation of traits through the correlation of trait expression with cDNA expression profiles. In spite of the frequent application of this method, studies on the optimization of the cDNA-AFLP assay design are rare and have typically been taxonomically restricted. Here, we model cDNA-AFLPs on all 92 eukaryotic species for which cDNA pools are currently available, using all combinations of eight restriction enzymes standard in cDNA-AFLP screens.ResultsIn silco simulations reveal that cDNA pool coverage is largely determined by the choice of individual restriction enzymes and that, through the choice of optimal enzyme combinations, coverage can be increased from <40% to 75% without changing the underlying experimental design. We find evidence of phylogenetic signal in the coverage data, which is largely mediated by organismal GC content. There is nonetheless a high degree of consistency in cDNA pool coverage for particular enzyme combinations, indicating that our recommendations should be applicable to most eukaryotic systems. We also explore the relationship between the average observed fragment number per selective AFLP-PCR reaction and the size of the underlying cDNA pool, and show how AFLP experiments can be used to estimate the number of genes expressed in a target tissue.ConclusionThe insights gained from in silico screening of cDNA-AFLPs from a broad sampling of eukaryotes provide a set of guidelines that should help to substantially increase the efficiency of future cDNA-AFLP experiments in eukaryotes. In silico simulations also suggest a novel use of cDNA-AFLP screens to determine the number of transcripts expressed in a target tissue, an application that should be invaluable as next-generation sequencing technologies are adapted for differential display.
| S-EPMC2797533 | biostudies-literature
Project description:Background. Effective treatment methods are needed for moderate/severely impairment chronic stroke. Objective. The questions were the following: (1) Is there need for long-dose therapy or is there a mid-treatment plateau? (2) Are the observed gains from the prior-studied protocol retained after treatment? Methods. Single-blind, stratified/randomized design, with 3 applied technology treatment groups, combined with motor learning, for long-duration treatment (300 hours of treatment). Measures were Arm Motor Ability Test time and coordination-function (AMAT-T, AMAT-F, respectively), acquired pre-/posttreatment and 3-month follow-up (3moF/U); Fugl-Meyer (FM), acquired similarly with addition of mid-treatment. Findings. There was no group difference in treatment response (P ≥ .16), therefore data were combined for remaining analyses (n = 31; except for FM pre/mid/post, n = 36). Pre-to-Mid-treatment and Mid-to-Posttreatment gains of FM were statistically and clinically significant (P < .0001; 4.7 points and P < .001; 5.1 points, respectively), indicating no plateau at 150 hours and benefit of second half of treatment. From baseline to 3moF/U: (1) FM gains were twice the clinically significant benchmark, (2) AMAT-F gains were greater than clinically significant benchmark, and (3) there was statistically significant improvement in FM (P < .0001); AMAT-F (P < .0001); AMAT-T (P < .0001). These gains indicate retained clinically and statistically significant gains at 3moFU. From posttreatment to 3moF/U, gains on FM were maintained. There were statistically significant gains in AMAT-F (P = .0379) and AMAT-T P = .003.
| S-EPMC6625035 | biostudies-literature
Project description:In a recidivism prediction context, there is no consensus on which modeling strategy should be followed for obtaining an optimal prediction model. In previous papers, a range of statistical and machine learning techniques were benchmarked on recidivism data with a binary outcome. However, two important tree ensemble methods, namely gradient boosting and random forests were not extensively evaluated. In this paper, we further explore the modeling potential of these techniques in the binary outcome criminal prediction context. Additionally, we explore the predictive potential of classical statistical and machine learning methods for censored time-to-event data. A range of statistical manually specified statistical and (semi-)automatic machine learning models is fitted on Dutch recidivism data, both for the binary outcome case and censored outcome case. To enhance generalizability of results, the same models are applied to two historical American data sets, the North Carolina prison data. For all datasets, (semi-) automatic modeling in the binary case seems to provide no improvement over an appropriately manually specified traditional statistical model. There is however evidence of slightly improved performance of gradient boosting in survival data. Results on the reconviction data from two sources suggest that both statistical and machine learning should be tried out for obtaining an optimal model. Even if a flexible black-box model does not improve upon the predictions of a manually specified model, it can serve as a test whether important interactions are missing or other misspecification of the model are present and can thus provide more security in the modeling process.
| S-EPMC6407787 | biostudies-literature
Project description:We report on the unusual, advantageous ageing of flexible transparent electrodes (FTEs) that were self-assembled from oleylamine-capped gold nanospheres (AuNPs) by direct nanoimprinting of inks with different particle concentrations (cAu = 3 mg mL-1 to 30 mg mL-1). The resulting lines were less than 2.5 μm wide and consisted of disordered particle assemblies. Small-Angle X-ray Scattering confirmed that particle packing did not change with ink concentration. Plasma sintering converted the printed structures into lines with a thin, electrically conductive metal shell and a less conductive hybrid core. We studied the opto-electronic performance directly after plasma sintering and after fourteen days of storage at 22 °C and 55% rH in the dark. The mean optical transmittance T̄400-800 in the range from 400 nm to 800 nm increased by up to ≈ 3%, while the sheet resistance Rsh strongly decreased by up to ≈ 82% at all concentrations. We correlated the changes with morphological changes visible in scanning and transmission electron microscopy and identified two sequential ageing stages: (I) post-plasma relaxation effects in and consolidation of the shell, and (II) particle re-organization, de-mixing, coarsening, and densification of the core with plating of Au from the core onto the shell, followed by solid-state de-wetting (ink concentrations cAu < 15 mg mL-1) or stability (cAu ≥ 15 mg mL-1). The plating of Au from the hybrid core improved the FTEs' Figure of Merit FOM = T̄400-800·Rsh-1 by up to ≈ 5.8 times and explains the stable value of ≈ 3.3%·Ωsq-1 reached after 7 days of ageing at cAu = 30 mg mL-1.
| S-EPMC10408613 | biostudies-literature