Search

Motivation In the realm of precision medicine, effective patient stratification and disease subtyping demand innovative methodologies tailored for multi-omics data. Clustering techniques applied to multi-omics data have become instrumental in identifying distinct subgroups of patients, enabling a finer-grained understanding of disease variability. Meanwhile, clinical datasets are often small and must be aggregated from multiple hospitals. Online data sharing, however, is seen as a significant challenge due to privacy concerns, potentially impeding big data's role in medical advancements using machine learning. This work establishes a powerful framework for advancing precision medicine through unsupervised random forest-based clustering in combination with federated computing. Results We introduce a novel multi-omics clustering approach utilizing unsupervised random forests. The unsupervised nature of the random forest enables the determination of cluster-specific feature importance, unraveling key molecular contributors to distinct patient groups. Our methodology is designed for federated execution, a crucial aspect in the medical domain where privacy concerns are paramount. We have validated our approach on machine learning benchmark datasets as well as on cancer data from The Cancer Genome Atlas. Our method is competitive with the state-of-the-art in terms of disease subtyping, but at the same time substantially improves the cluster interpretability. Experiments indicate that local clustering performance can be improved through federated computing. Availability and implementation The proposed methods are available as an R-package (https://github.com/pievos101/uRF) [ABSTRACT FROM AUTHOR]

In our previous work, we introduced a method that combines two unsupervised algorithms: Density-based spatial clustering of applications with noise (DBSCAN) and Gaussian mixture model (GMM). We applied this method to 12 open clusters based on Gaia Early Data Release 3 (EDR3) data, demonstrating its effectiveness in identifying reliable cluster members within the tidal radius. However, for studying cluster morphology, we need a method capable of detecting members both inside and outside the tidal radius. By incorporating a supervised algorithm into our approach, we successfully identified members beyond the tidal radius. In our current work, we initially applied DBSCAN and GMM to identify reliable members of cluster stars. Subsequently, we trained the random forest algorithm using DBSCAN- and GMM-selected data. Leveraging the random forest, we can identify cluster members outside the tidal radius and observe cluster morphology across a wide field of view. Our method was then applied to 15 open clusters based on Gaia DR3, which exhibit a wide range of metallicity, distances, members, and ages. Additionally, we calculated the tidal radius for each of the 15 clusters using the King profile and detected stars both inside and outside this radius. Finally, we investigated mass segregation and luminosity distribution within the clusters. Overall, our approach significantly improved the estimation of the tidal radius and detection of mass segregation compared to the previous work. We found that in Collinder 463, low-mass stars do not segregate in comparison to high-mass and intermediate-mass stars. Additionally, we detected a peak of luminosity in the clusters, some of which were located far from the centre, beyond the tidal radius. [ABSTRACT FROM AUTHOR]

Motivation High-throughput RNA sequencing has become indispensable for decoding gene activities, yet the challenge of reconstructing full-length transcripts persists. Traditional single-sample assemblers frequently produce fragmented transcripts, especially in single-cell RNA-seq data. While algorithms designed for assembling multiple samples exist, they encounter various limitations. Results We present Aletsch, a new assembler for multiple bulk or single-cell RNA-seq samples. Aletsch incorporates several algorithmic innovations, including a "bridging" system that can effectively integrate multiple samples to restore missed junctions in individual samples, and a new graph-decomposition algorithm that leverages "supporting" information across multiple samples to guide the decomposition of complex vertices. A standout feature of Aletsch is its application of a random forest model with 50 well-designed features for scoring transcripts. We demonstrate its robust adaptability across different chromosomes, datasets, and species. Our experiments, conducted on RNA-seq data from several protocols, firmly demonstrate Aletsch's significant outperformance over existing meta-assemblers. As an example, when measured with the partial area under the precision-recall curve (pAUC, constrained by precision), Aletsch surpasses the leading assemblers TransMeta by 22.9%–62.1% and PsiCLASS by 23.0%–175.5% on human datasets. Availability and implementation Aletsch is freely available at https://github.com/Shao-Group/aletsch. Scripts that reproduce the experimental results of this manuscript is available at https://github.com/Shao-Group/aletsch-test. [ABSTRACT FROM AUTHOR]

Motivation Understanding the mechanisms and structural mappings between molecules and pathway classes are critical for design of reaction predictors for synthesizing new molecules. This article studies the problem of prediction of classes of metabolic pathways (series of chemical reactions occurring within a cell) in which a given biochemical compound participates. We apply a hybrid machine learning approach consisting of graph convolutional networks used to extract molecular shape features as input to a random forest classifier. In contrast to previously applied machine learning methods for this problem, our framework automatically extracts relevant shape features directly from input SMILES representations, which are atom-bond specifications of chemical structures composing the molecules. Results Our method is capable of correctly predicting the respective metabolic pathway class of 95.16% of tested compounds, whereas competing methods only achieve an accuracy of 84.92% or less. Furthermore, our framework extends to the task of classification of compounds having mixed membership in multiple pathway classes. Our prediction accuracy for this multi-label task is 95.62%. We analyze the relative importance of various global physicochemical features to the pathway class prediction problem and show that simple linear/logistic regression models can predict the values of these global features from the shape features extracted using our framework. Availability and implementation https://github.com/baranwa2/MetabolicPathwayPrediction. [ABSTRACT FROM AUTHOR]

Bycatch rates are essential to estimating fishery impacts and making management decisions, but data on bycatch are often limited. Tropical tuna purse seine (PS) fisheries catch numerous bycatch species, including vulnerable silky sharks. Even if bycatch proportion is relatively low, impacts on pelagic ecosystems may be important due to the large size of these fisheries. Partial observer coverage of bycatch is a major impediment to assessing impacts. Here we develop a generic Δ modeling approach for predicting catch of four major bycatch species, including silky sharks, in floating object-associated fishing sets of the French Indian Ocean PS fleet from 2011 to 2018 based on logbook and observer data. Cross-validation and variable selection are used to identify optimal models consisting of a random forest model for presence–absence and a negative binomial general-additive model for abundance when present. Though models explain small to moderate amounts of variance (5–15%), they outperform a simpler approach commonly used for reporting, and they allow us to estimate total annual bycatch for the four species with robust estimates of uncertainty. Interestingly, uncertainty relative to mean catch is lower for top predators than forage species, consistent with these species having similar behavior and ecological niches to tunas. [ABSTRACT FROM AUTHOR]

Studies have identified genes and molecular pathways regulating cancer metastasis. However, it remains largely unknown whether metastatic potentials of cancer cells from different lineage types are driven by the same or different gene networks. Here, we aim to address this question through integrative analyses of 493 human cancer cells' transcriptomic profiles and their metastatic potentials in vivo. Using an unsupervised approach and considering both gene coexpression and protein–protein interaction networks, we identify different gene networks associated with various biological pathways (i.e. inflammation, cell cycle, and RNA translation), the expression of which are correlated with metastatic potentials across subsets of lineage types. By developing a regularized random forest regression model, we show that the combination of the gene module features expressed in the native cancer cells can predict their metastatic potentials with an overall Pearson correlation coefficient of 0.90. By analyzing transcriptomic profile data from cancer patients, we show that these networks are conserved in vivo and contribute to cancer aggressiveness. The intrinsic expression levels of these networks are correlated with drug sensitivity. Altogether, our study provides novel comparative insights into cancer cells' intrinsic gene networks mediating metastatic potentials across different lineage types, and our results can potentially be useful for designing personalized treatments for metastatic cancers. [ABSTRACT FROM AUTHOR]

Recent advances in sequencing, mass spectrometry, and cytometry technologies have enabled researchers to collect multiple 'omics data types from a single sample. These large datasets have led to a growing consensus that a holistic approach is needed to identify new candidate biomarkers and unveil mechanisms underlying disease etiology, a key to precision medicine. While many reviews and benchmarks have been conducted on unsupervised approaches, their supervised counterparts have received less attention in the literature and no gold standard has emerged yet. In this work, we present a thorough comparison of a selection of six methods, representative of the main families of intermediate integrative approaches (matrix factorization, multiple kernel methods, ensemble learning, and graph-based methods). As non-integrative control, random forest was performed on concatenated and separated data types. Methods were evaluated for classification performance on both simulated and real-world datasets, the latter being carefully selected to cover different medical applications (infectious diseases, oncology, and vaccines) and data modalities. A total of 15 simulation scenarios were designed from the real-world datasets to explore a large and realistic parameter space (e.g. sample size, dimensionality, class imbalance, effect size). On real data, the method comparison showed that integrative approaches performed better or equally well than their non-integrative counterpart. By contrast, DIABLO and the four random forest alternatives outperform the others across the majority of simulation scenarios. The strengths and limitations of these methods are discussed in detail as well as guidelines for future applications. [ABSTRACT FROM AUTHOR]

A major challenge of precision oncology is the identification and prioritization of suitable treatment options based on molecular biomarkers of the considered tumor. In pursuit of this goal, large cancer cell line panels have successfully been studied to elucidate the relationship between cellular features and treatment response. Due to the high dimensionality of these datasets, machine learning (ML) is commonly used for their analysis. However, choosing a suitable algorithm and set of input features can be challenging. We performed a comprehensive benchmarking of ML methods and dimension reduction (DR) techniques for predicting drug response metrics. Using the Genomics of Drug Sensitivity in Cancer cell line panel, we trained random forests, neural networks, boosting trees and elastic nets for 179 anti-cancer compounds with feature sets derived from nine DR approaches. We compare the results regarding statistical performance, runtime and interpretability. Additionally, we provide strategies for assessing model performance compared with a simple baseline model and measuring the trade-off between models of different complexity. Lastly, we show that complex ML models benefit from using an optimized DR strategy, and that standard models—even when using considerably fewer features—can still be superior in performance. [ABSTRACT FROM AUTHOR]

Runs of homozygosity (ROHs) are indicative of elevated homozygosity and inbreeding due to mating of closely related individuals. Self-fertilization can be a major source of inbreeding which elevates genome-wide homozygosity and thus should also create long ROHs. While ROHs are frequently used to understand inbreeding in the context of conservation and selective breeding, as well as for consanguinity of populations and their demographic history, it remains unclear how ROH characteristics are altered by selfing and if this confounds expected signatures of inbreeding due to demographic change. Using simulations, we study the impact of the mode of reproduction and demographic history on ROHs. We apply random forests to identify unique characteristics of ROHs, indicative of different sources of inbreeding. We pinpoint distinct features of ROHs that can be used to better characterize the type of inbreeding the population was subjected to and to predict outcrossing rates and complex demographic histories. Using additional simulations and four empirical datasets, two from highly selfing species and two from mixed-maters, we predict the selfing rate and validate our estimations. We find that self-fertilization rates are successfully identified even with complex demography. Population genetic summary statistics improve algorithm accuracy particularly in the presence of additional inbreeding, e.g. from population bottlenecks. Our findings highlight the importance of ROHs in disentangling confounding factors related to various sources of inbreeding and demonstrate situations where such sources cannot be differentiated. Additionally, our random forest models provide a novel tool to the community for inferring selfing rates using genomic data. [ABSTRACT FROM AUTHOR]

Small fractions of patients suffer from radiotherapy late severe adverse events (AEs Grade ≥ 3), which are usually irreversible and badly affect their quality of life. A novel functional DNA repair assay characterizing several steps of double-strand break (DSB) repair mechanisms was used. DNA repair activities of peripheral blood mononuclear cells were monitored for 1 week using NEXT-SPOT assay in 177 breast and prostate cancer patients. Only seven patients had Grade ≥ 3 AEs, 6 months after radiotherapy initiation. The machine learning method established the importance of variables among demographic, clinical and DNA repair data. The most relevant ones, all related to DNA repair, were employed to build a predictor. Predictors constructed with random forest and minimum bounding sphere predicted late Grade ≥ 3 AEs with a sensitivity of 100% and specificity of 77.17 and 86.22%, respectively. This multiplex functional approach strongly supports a dominant role for DSB repair in the development of chronic AEs. It also showed that affected patients share specific features related to functional aspects of DSB repair. This strategy may be suitable for routine clinical analysis and paves the way for modelling DSB repair associated with severe AEs induced by radiotherapy. [ABSTRACT FROM AUTHOR]

The advent of massive broad-band photometric surveys enabled photometric redshift estimates for unprecedented numbers of galaxies and quasars. These estimates can be improved using better algorithms or by obtaining complementary data such as narrow-band photometry, and broad-band photometry over an extended wavelength range. We investigate the impact of both approaches on photometric redshifts for quasars using data from Southern Photometric Local Universe Survey (S-PLUS) DR4, Galaxy Evolution Explorer (GALEX) DR6/7, and the unWISE catalog for the Wide-field Infrared Survey Explorer (WISE) in three machine learning methods: Random Forest, Flexible Conditional Density Estimation (FlexCoDE), and Bayesian Mixture Density Network (BMDN). Including narrow-band photometry improves the root-mean-square error by 11 per cent in comparison to a model trained with only broad-band photometry. Narrow-band information only provided an improvement of 3.8 per cent when GALEX and WISE colours were included. Thus, narrow bands play a more important role for objects that do not have GALEX or WISE counterparts, which respectively makes 92 per cent and 25 per cent of S-PLUS data considered here. Nevertheless, the inclusion of narrow-band information provided better estimates of the probability density functions obtained with FlexCoDE and BMDN. We publicly release a value-added catalogue of photometrically selected quasars with the photo- z predictions from all methods studied here. The catalogue provided with this work covers the S-PLUS DR4 area (∼3000 square degrees), containing 645 980, 244 912, 144 991 sources with the probability of being a quasar higher than, 80 per cent, 90 per cent, 95 per cent up to r < 21.3 and good photometry quality in the detection image. More quasar candidates can be retrieved from the S-PLUS data base by considering less restrictive selection criteria. [ABSTRACT FROM AUTHOR]

Pseudouridine is an RNA modification that is widely distributed in both prokaryotes and eukaryotes, and plays a critical role in numerous biological activities. Despite its importance, the precise identification of pseudouridine sites through experimental approaches poses significant challenges, requiring substantial time and resources.Therefore, there is a growing need for computational techniques that can reliably and quickly identify pseudouridine sites from vast amounts of RNA sequencing data. In this study, we propose fuzzy kernel evidence Random Forest (FKeERF) to identify pseudouridine sites. This method is called PseU-FKeERF, which demonstrates high accuracy in identifying pseudouridine sites from RNA sequencing data. The PseU-FKeERF model selected four RNA feature coding schemes with relatively good performance for feature combination, and then input them into the newly proposed FKeERF method for category prediction. FKeERF not only uses fuzzy logic to expand the original feature space, but also combines kernel methods that are easy to interpret in general for category prediction. Both cross-validation tests and independent tests on benchmark datasets have shown that PseU-FKeERF has better predictive performance than several state-of-the-art methods. This new method not only improves the accuracy of pseudouridine site identification, but also provides a certain reference for disease control and related drug development in the future. [ABSTRACT FROM AUTHOR]

This study examines the feature space of seismic waveforms often used in machine learning applications for seismic event detection and classification problems. Our investigation centres on the southern Alaska region, where the seismic record captures diverse seismic activity, notably from the calving of marine-terminating glaciers and tectonic earthquakes along active plate boundaries. While the automated discrimination of earthquakes and glacier quakes is our nominal goal, this data set provides an outstanding opportunity to explore the general feature space of regional seismic phases. That objective has applicability beyond ice quakes and our geographic region of study. We make a noteworthy discovery that features rooted in the spectral content of seismic waveforms consistently outperform statistical and temporal features. Spectral features demonstrate robust performance, exhibiting resilience to class imbalance while being minimally impacted by factors such as epicentral distance and signal-to-noise ratio. We also conduct experiments on the transferability of the model and find that transferability primarily depends on the appearance of the waveforms. Finally, we analyse misclassified events and find examples that are identified incorrectly in the original regional catalogue. [ABSTRACT FROM AUTHOR]

While the available set of gamma-ray burst (GRB) data with known redshift is currently limited, a much larger set of GRB data without redshift is available from different instruments. This data includes well-measured prompt gamma-ray flux and spectral information. We estimate the redshift of a selection of these GRBs detected by Fermi -GBM and Konus- Wind using machine learning techniques that are based on spectral parameters. We find that Deep Neural Networks with Random Forest models employing non-linear relations among input parameters can reasonably reproduce the pseudo-redshift distribution of GRBs, mimicking the distribution of GRBs with spectroscopic redshift. Furthermore, we find that the pseudo-redshift samples of GRBs to satisfy (i) Amati relation between the peak photon energy of the time-averaged energy spectrum in the cosmological rest frame of the GRB E i,p and the isotropic-equivalent radiated energy E iso during the prompt phase; and (ii) Yonetoku relation between E i,p and isotropic-equivalent luminosity L iso, both measured during the peak flux interval. [ABSTRACT FROM AUTHOR]

Diabetes prevalence is rising globally, especially in low- and middle-income countries like Mexico, posing challenges for healthcare systems that require efficient primary care to manage the disease. However, healthcare efficiency is influenced by factors beyond decision-makers, including socioeconomic and political conditions. This study aims to evaluate the technical efficiency of primary healthcare for diabetes patients in Mexico over a 12-year period and explore the impact of contextual variables on efficiency. A longitudinal analysis was conducted using administrative and socio-demographic data from 242 health jurisdictions between 2009 and 2020. Data envelopment analysis with bootstrapping and output orientation was used to measure the technical efficiency; health resources in infrastructure and human resources were used as inputs. As outcome, the number of patients receiving treatment for diabetes and the number of patients with controlled diabetes were considered. Machine learning algorithms were employed to analyse multiple factors affecting the provision of diabetes health services and assess heterogeneity and trends in efficiency across different health jurisdictions. The average technical efficiency in primary healthcare for diabetes patients was 0.44 (CI: 0.41–0.46) in 2009, reaching a peak of 0.71 (CI: 0.69–0.72) in 2016, and moderately declining to 0.60 (CI: 0.57–0.62) in 2020; these differences were statistically significant. The random forest analysis identified the marginalization index, primary healthcare coverage, proportion of indigenous population and demand for health services as the most influential variables in predicting efficiency levels. This research underscores the crucial need for the formulation of targeted public policies aimed at extending the scope of primary healthcare services, with a particular focus on addressing the unique challenges faced by marginalized and indigenous populations. According to our results, it is necessary that medical care management adjust to the specific demands and needs of these populations to guarantee equitable care in Mexico. [ABSTRACT FROM AUTHOR]

Non-parametric morphology statistics have been used for decades to classify galaxies into morphological types and identify mergers in an automated way. In this work, we assess how reliably we can identify galaxy post-mergers with non-parametric morphology statistics. Low-redshift (z ≲ 0.2), recent (t post-merger ≲ 200 Myr), and isolated (r > 100 kpc) post-merger galaxies are drawn from the IllustrisTNG100-1 cosmological simulation. Synthetic r -band images of the mergers are generated with SKIRT9 and degraded to various image qualities, adding observational effects such as sky noise and atmospheric blurring. We find that even in perfect quality imaging, the individual non-parametric morphology statistics fail to recover more than 55 per cent of the post-mergers, and that this number decreases precipitously with worsening image qualities. The realistic distributions of galaxy properties in IllustrisTNG allow us to show that merger samples assembled using individual morphology statistics are biased towards low-mass, high gas fraction, and high mass ratio. However, combining all of the morphology statistics together using either a linear discriminant analysis or random forest algorithm increases the completeness and purity of the identified merger samples and mitigates bias with various galaxy properties. For example, we show that in imaging similar to that of the 10-yr depth of the Legacy Survey of Space and Time, a random forest can identify 89 per cent of mergers with a false positive rate of 17 per cent. Finally, we conduct a detailed study of the effect of viewing angle on merger observability and find that there may be an upper limit to merger recovery due to the orientation of merger features with respect to the observer. [ABSTRACT FROM AUTHOR]

Motivation Random forests (RFs) can deal with a large number of variables, achieve reasonable prediction scores, and yield highly interpretable feature importance values. As such, RFs are appropriate models for feature selection and further dimension reduction. However, RFs are often not appropriate for correlated datasets due to their mode of selecting individual features for splitting. Addressing correlation relationships in high-dimensional datasets is imperative for reducing the number of variables that are assigned high importance, hence making the dimension reduction most efficient. Here, we propose the LAtent VAriable Stochastic Ensemble of Trees (LAVASET) method that derives latent variables based on the distance characteristics of each feature and aims to incorporate the correlation factor in the splitting step. Results Without compromising on performance in the majority of examples, LAVASET outperforms RF by accurately determining feature importance across all correlated variables and ensuring proper distribution of importance values. LAVASET yields mostly non-inferior prediction accuracies to traditional RFs when tested in simulated and real 1D datasets, as well as more complex and high-dimensional 3D datatypes. Unlike traditional RFs, LAVASET is unaffected by single 'important' noisy features (false positives), as it considers the local neighbourhood. LAVASET, therefore, highlights neighbourhoods of features, reflecting real signals that collectively impact the model's predictive ability. Availability and implementation LAVASET is freely available as a standalone package from https://github.com/melkasapi/LAVASET. [ABSTRACT FROM AUTHOR]

Metabolomics and foodomics shed light on the molecular processes within living organisms and the complex food composition by leveraging sophisticated analytical techniques to systematically analyze the vast array of molecular features. The traditional feature-picking method often results in arbitrary selections of the model, feature ranking, and cut-off, which may lead to suboptimal results. Thus, a Multiple and Optimal Screening Subset (MOSS) approach was developed in this study to achieve a balance between a minimal number of predictors and high predictive accuracy during statistical model setup. The MOSS approach compares five commonly used models in the context of food matrix analysis, specifically bourbons. These models include Student's t -test, receiver operating characteristic curve, partial least squares-discriminant analysis (PLS-DA), random forests, and support vector machines. The approach employs cross-validation to identify promising subset feature candidates that contribute to food characteristic classification. It then determines the optimal subset size by comparing it to the corresponding top-ranked features. Finally, it selects the optimal feature subset by traversing all possible feature candidate combinations. By utilizing MOSS approach to analyze 1406 mass spectral features from a collection of 122 bourbon samples, we were able to generate a subset of features for bourbon age prediction with 88% accuracy. Additionally, MOSS increased the area under the curve performance of sweetness prediction to 0.898 with only four predictors compared with the top-ranked four features at 0.681 based on the PLS-DA model. Overall, we demonstrated that MOSS provides an efficient and effective approach for selecting optimal features compared with other frequently utilized methods. [ABSTRACT FROM AUTHOR]

As the application of large language models (LLMs) has broadened into the realm of biological predictions, leveraging their capacity for self-supervised learning to create feature representations of amino acid sequences, these models have set a new benchmark in tackling downstream challenges, such as subcellular localization. However, previous studies have primarily focused on either the structural design of models or differing strategies for fine-tuning, largely overlooking investigations into the nature of the features derived from LLMs. In this research, we propose different ESM2 representation extraction strategies, considering both the character type and position within the ESM2 input sequence. Using model dimensionality reduction, predictive analysis and interpretability techniques, we have illuminated potential associations between diverse feature types and specific subcellular localizations. Particularly, the prediction of Mitochondrion and Golgi apparatus prefer segments feature closer to the N-terminal, and phosphorylation site-based features could mirror phosphorylation properties. We also evaluate the prediction performance and interpretability robustness of Random Forest and Deep Neural Networks with varied feature inputs. This work offers novel insights into maximizing LLMs' utility, understanding their mechanisms, and extracting biological domain knowledge. Furthermore, we have made the code, feature extraction API, and all relevant materials available at https://github.com/yujuan-zhang/feature-representation-for-LLMs. [ABSTRACT FROM AUTHOR]

Bullet ricochets are common occurrences during shooting incidents and can provide a wealth of information useful for shooting incident reconstruction. However, there have only been a small number of studies that have systematically investigated bullet ricochet impact site morphology. Here, this study reports on an experiment that examined the plan-view morphology of 297 ricochet impact sites in concrete that were produced by five different bullet types shot from two distances. This study used a random forest machine learning algorithm to classify bullet types with morphological dimensions of the ricochet mark (impact) with length and perimeter-to-area ratio emerging as the top predictor variables. The 0.22 LR leaves the most distinctive impact mark on the concrete, and overall, the classification accuracy using leave-one-out cross-validation is 62%, considerably higher than a random classification accuracy of 20%. Adding in distance to the model as a predictor increases the classification accuracy to 66%. These initial results are promising, in that they suggest that an unknown bullet type can potentially be determined, or at least probabilistically assessed, from the morphology of the ricochet impact site alone. However, the substantial amount of overlap this study documented among distinct bullet types' ricochet mark morphologies under highly controlled conditions and with machine learning suggests that the human identification of ricochet marks in real-world shooting incident reconstructions may be on occasion, or perhaps regularly, in error. Key points Bullet ricochet impact sites can help with shooting incident reconstruction. A random forest machine learning algorithm classified bullet type from ricochet morphology. Results suggest that unknown bullets can potentially be determined from ricochet impact site morphology. Human identification of bullet types from ricochet sites may be erroneous. [ABSTRACT FROM AUTHOR]

Accurate estimation of photometric redshifts (photo- z) is crucial in studies of both galaxy evolution and cosmology using current and future large sky surveys. In this study, we employ Random Forest (RF), a machine learning algorithm, to estimate photo- z and investigate the systematic uncertainties affecting the results. Using galaxy flux and colour as input features, we construct a mapping between input features and redshift by using a training set of simulated data, generated from the Hubble Space Telescope Advanced Camera for Surveys (HST-ACS) and COSMOS catalogue, with the expected instrumental effects of the planned China Space Station Telescope (CSST). To improve the accuracy and confidence of predictions, we incorporate inverse variance weighting and perturb the catalogue using input feature errors. Our results show that weighted RF can achieve a photo- z accuracy of |$\rm \sigma _{NMAD}=0.025$| and an outlier fraction of |$\rm \eta =2.045\ \hbox{per cent}$|⁠ , significantly better than the values of |$\rm \sigma _{NMAD}=0.043$| and |$\rm \eta =6.45\ \hbox{per cent}$| obtained by the widely used Easy and Accurate Zphot from Yale (EAZY) software, which uses template-fitting method. Furthermore, we have calculated the importance of each input feature for different redshift ranges and found that the most important input features reflect the approximate position of the break features in galaxy spectra, demonstrating the algorithm's ability to extract physical information from data. Additionally, we have established confidence indices and error bars for each prediction value based on the shape of the redshift probability distribution function, suggesting that screening sources with high confidence can further reduce the outlier fraction. [ABSTRACT FROM AUTHOR]

We present the extended ALMA MaNGA QUEnching and STar formation survey (ALMaQUEST), a combination of the original 46 ALMaQUEST galaxies plus new ALMA observations for a further 20 interacting galaxies. Three well-studied scaling relations are fit to the 19 999 star-forming spaxels in the extended sample, namely the resolved Schmidt–Kennicutt relation, the resolved star-forming main-sequence and the resolved molecular gas main sequence. We additionally investigate the relationship between the dynamical equilibrium pressure (P DE) and star formation rate surface density (ΣSFR), which we refer to as the resolved P DE (rPDE) relation. Contrary to previous studies that have focussed on normal star-forming galaxies and found an approximately linear rPDE relation, the presence of more vigourously star-forming galaxies in the extended ALMaQUEST sample reveals a marked turnover in the relation at high pressures. Although the scatter around the linear fit to the rPDE relation is similar to the other three relations, a random forest analysis, which can extract non-linear dependences, finds that P DE is unambiguously more important than either |$\Sigma _{\rm H_2}$| or Σ⋆ for predicting Σ SFR. We compare the observed rPDE relation to the prediction of the pressure-regulated feedback-modulated (PRFM) model of star formation, finding that galaxies residing on the global SFMS do indeed closely follow the rPDE relation predicted by the PRFM theory. However, galaxies above and below the global SFMS show significant deviations from the model. Galaxies with high SFR are instead consistent with models that include other contributions to turbulence in addition to the local star formation feedback. [ABSTRACT FROM AUTHOR]

Polar regions and Greenland in particular are highly sensitive to global warming. Impacts on Greenland's glaciers may be observed through the increasing number of calving events. However, a direct assessment of the calving activity is limited due to the remoteness of polar regions and the cloudy weather which makes impossible a recurrent observation through satellite imagery. To tackle this issue, we exploit the seismological network deployed in Greenland which actively records seismic signals associated with calving events, hereinafter referred to as glacial earthquakes. These seismic signals present a broad frequency range and a wide diversity of waveform which make them difficult to discriminate from tectonic events as well as anthropogenic and natural noises. In this study, we start from two catalogues of known events, one for glacial earthquake events which occurred between 1993 and 2013 and one for earthquakes which occurred in the same time period, and we implement a detection algorithm based on the STA/LTA method to extract signals' events from continuous data. Then, we train and test a machine learning processing chain based on the Random Forest algorithm which allows us to automatically associate the events respectively with calving and tectonic activity, with a certain probability. Finally, we investigate 844 selected days spanning time of continuous data from the Greenland regional seismic network which results in a new, more exhaustive, catalogue of glacial earthquakes expanded of 1633 newly detected glacial events. Moreover, we extensively discuss the choice of the features used to describe glacial earthquakes, in particular the 39 new features created in this study which have drastically improved our results with 7 of the 10 best features being in the added set. The perspective of further expansion of the glacial earthquake catalogue applying the processing chain discussed in this paper on different time spans highlights how combining seismology and machine learning can increase our understanding of the spatio-temporal evolution of calving activity in remote regions. [ABSTRACT FROM AUTHOR]

Background: Erectile dysfunction (ED) is a common condition affecting middle-aged and elderly men. Aim: The study sought to investigate differentially expressed fatty acid metabolism–related genes and the molecular mechanisms of ED. Methods: The expression profiles of GSE2457 and GSE31247 were downloaded from the Gene Expression Omnibus database and merged. Differentially expressed genes (DEGs) between ED and normal samples were obtained using the R package limma. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses of DEGs were conducted using the R package clusterProfiler. Fatty acid metabolism–related DEGs (FAMDEGs) were further identified and analyzed. Machine learning algorithms, including Lasso (least absolute shrinkage and selection operator), support vector machine, and random forest algorithms, were utilized to identify hub FAMDEGs with the ability to predict ED occurrence. Coexpression analysis and gene set enrichment analysis of hub FAMDEGs were performed. Outcome: Fatty acid metabolism–related functions (such as fatty acid metabolism and degradation) may play a vital role in ED. Results: In total, 5 hub FAMDEGs (Aldh2, Eci2, Acat1, Acadl, and Hadha) were identified and found to be differentially expressed between ED and normal samples. Gene set enrichment analysis identified key pathways associated with these genes. The area under the curve values of the 5 hub FAMDEGs for predicting ED occurrence were all >0.8. Clinical Translation: Our results suggest that these 5 key FAMDEGs may serve as biomarkers for the diagnosis and treatment of ED. Strengths and Limitations: The strengths of our study include the use of multiple datasets and machine learning algorithms to identify key FAMDEGs. However, limitations include the lack of validation in animal models and human tissues, as well as research on the mechanisms of these FAMDEGs. Conclusion: Five hub FAMDEGs were identified as potential biomarkers for ED progression. Our work may prove that fatty acid metabolism–related genes are worth further investigation in ED. Graphical Abstract [ABSTRACT FROM AUTHOR]

We explore the dependence of dust attenuation, as traced by the |$\rm H_{\alpha }/\rm H_{\beta }$| Balmer decrement, on galactic properties by using a large sample of Sloan Digital Sky Survey spectra. We use both partial correlation coefficients and random forest analysis to distinguish those galactic parameters that directly and primarily drive dust attenuation in galaxies, from parameters that are only indirectly correlated through secondary dependencies. We find that, once galactic inclination is controlled for, dust attenuation depends primarily on stellar mass, followed by metallicity and velocity dispersion. Once the dependence on these quantities is taken into account, there is no dependence on the star formation rate. While the dependence on stellar mass and metallicity was expected based on simple analytical equations for the interstellar medium, the dependence on velocity dispersion was not predicted, and we discuss possible scenarios to explain it. We identify a projection of this multidimensional parameters space which minimizes the dispersion in terms of the Balmer decrement and which encapsulates the primary and secondary dependences of the Balmer decrement into a single parameter defined as the reduced mass μ = log M⋆ + 3.67[O/H] + 2.96log (σ v /100 km s−1). We show that the dependence of the Balmer decrement on this single parameter also holds at high redshift, suggesting that the processes regulating dust production and distribution do not change significantly through cosmic epochs at least out to z ∼ 2. [ABSTRACT FROM AUTHOR]

We present results of recurrence analysis of the black hole X-ray binary Cygnus X-1 using combined observations from the Rossi X-ray Timing Explorer All-sky Monitor and the Japanese Monitor of All-sky X-ray Image aboard the International Space Station. From the time-dependent windowed recurrence plot (RP), we compute 10 recurrence quantities that describe the dynamical behaviour of the source and compare them to the spectral state at each point in time. We identify epochs of state changes corresponding to transitions into highly deterministic or highly stochastic dynamical regimes and their correlation to specific spectral states. We compare k-Nearest Neighbors and Random Forest models for various sizes of the time-dependent RP. The spectral state in Cygnus X-1 can be predicted with greater than 95 per cent accuracy for both types of models explored across a range of RP sizes based solely on the recurrence properties. The primary features from the RP that distinguish between spectral states are the determinism, Shannon entropy, and average line length, all of which are systematically higher in the hard state compared to the soft state. Our results suggest that the hard and soft states of Cygnus X-1 exhibit distinct dynamical variability and the time domain alone can be used for spectral state classification. [ABSTRACT FROM AUTHOR]

We explore the feasibility of learning the connection between Sloan Digital Sky Survey (SDSS) galaxies and ELUCID subhaloes with random forest (RF). ELUCID is a constrained N -body simulation constructed using the matter density field of SDSS. Based on a SDSS-ELUCID matched catalogue, we build RF models that predict Mr magnitude, colour, stellar mass M *, and specific star formation rate (sSFR) with several subhalo properties. While the RF can predict Mr and M * with reasonable accuracy, the prediction accuracy of colour and sSFR is low, which could be due to the mismatch between galaxies and subhaloes. To test this, we shuffle the galaxies in subhaloes of narrow mass bins in the local neighbourhood using galaxies of a semi-analytic model (SAM) and the TNG hydrodynamic simulation. We find that the shuffling only slightly reduces the colour prediction accuracy in SAM and TNG, which is still considerably higher than that of the SDSS. This suggests that the true connection between SDSS colour and subhalo properties could be weaker than that in the SAM and TNG without the mismatch effect. We also measure the Pearson correlation coefficient between the galaxy and subhalo properties in SDSS, SAM, and TNG. Similar to the RF results, we find that the colour–subhalo correlation in SDSS is lower than in both SAM and TNG. We also show that the galaxy–subhalo correlations depend on subhalo mass in the galaxy formation models. Advanced surveys with fainter galaxies will provide new insights into the galaxy–subhalo relation in the real Universe. [ABSTRACT FROM AUTHOR]

This paper explores the application of machine learning methods for classifying astronomical sources using photometric data, including normal and emission line galaxies (ELGs; starforming, starburst, AGN, broad-line), quasars, and stars. We utilized samples from Sloan Digital Sky Survey (SDSS) Data Release 17 (DR17) and the ALLWISE catalogue, which contain spectroscopically labelled sources from SDSS. Our methodology comprises two parts. First, we conducted experiments, including three-class, four-class, and seven-class classifications, employing the Random Forest (RF) algorithm. This phase aimed to achieve optimal performance with balanced data sets. In the second part, we trained various machine learning methods, such as k -nearest neighbours (KNN), RF, XGBoost (XGB), voting, and artificial neural network (ANN), using all available data based on promising results from the first phase. Our results highlight the effectiveness of combining optical and infrared features, yielding the best performance across all classifiers. Specifically, in the three-class experiment, RF and XGB algorithms achieved identical average F1 scores of 98.93 per cent on both balanced and unbalanced data sets. In the seven-class experiment, our average F1 score was 73.57 per cent. Using the XGB method in the four-class experiment, we achieved F1 scores of 87.9 per cent for normal galaxies (NGs), 81.5 per cent for ELGs, 99.1 per cent for stars, and 98.5 per cent for quasars (QSOs). Unlike classical methods based on time-consuming spectroscopy, our experiments demonstrate the feasibility of using automated algorithms on carefully classified photometric data. With more data and ample training samples, detailed photometric classification becomes possible, aiding in the selection of follow-up observation candidates. [ABSTRACT FROM AUTHOR]

DNA methylation profiling is a useful tool to increase the accuracy of a cancer diagnosis. However, a comprehensive R package specially for it is lacking. Hence, we developed the R package methylClass for methylation-based classification. Within it, we provide the eSVM (ensemble-based support vector machine) model to achieve much higher accuracy in methylation data classification than the popular random forest model and overcome the time-consuming problem of the traditional SVM. In addition, some novel feature selection methods are included in the package to improve the classification. Furthermore, because methylation data can be converted to other omics, such as copy number variation data, we also provide functions for multi-omics studies. The testing of this package on four datasets shows the accurate performance of our package, especially eSVM, which can be used in both methylation and multi-omics models and outperforms other methods in both cases. methylClass is available at: https://github.com/yuabrahamliu/methylClass. [ABSTRACT FROM AUTHOR]

Throughout evolution, pathogenic viruses have developed different strategies to evade the response of the adaptive immune system. To carry out successful replication, some pathogenic viruses encode different proteins that manipulate the molecular mechanisms of host cells. Currently, there are different bioinformatics tools for virus research; however, none of them focus on predicting viral proteins that evade the adaptive system. In this work, we have developed a novel tool based on machine and deep learning for predicting this type of viral protein named VirusHound-I. This tool is based on a model developed with the multilayer perceptron algorithm using the dipeptide composition molecular descriptor. In this study, we have also demonstrated the robustness of our strategy for data augmentation of the positive dataset based on generative antagonistic networks. During the 10-fold cross-validation step in the training dataset, the predictive model showed 0.947 accuracy, 0.994 precision, 0.943 F1 score, 0.995 specificity, 0.896 sensitivity, 0.894 kappa, 0.898 Matthew's correlation coefficient and 0.989 AUC. On the other hand, during the testing step, the model showed 0.964 accuracy, 1.0 precision, 0.967 F1 score, 1.0 specificity, 0.936 sensitivity, 0.929 kappa, 0.931 Matthew's correlation coefficient and 1.0 AUC. Taking this model into account, we have developed a tool called VirusHound-I that makes it possible to predict viral proteins that evade the host's adaptive immune system. We believe that VirusHound-I can be very useful in accelerating studies on the molecular mechanisms of evasion of pathogenic viruses, as well as in the discovery of therapeutic targets. [ABSTRACT FROM AUTHOR]

The laboratory-derived rate- and state-dependent friction (RSF) law governs rock friction. Although a number of studies have investigated the RSF friction parameters, they are not fully understood yet within a physical framework. In this study, we estimated the variation of RSF parameters during stick-slip cycles, in order to have insights into the temporal variation of fault conditions during slipping, which may help understand the relation between the change in friction parameters and the generation of gouge particles. To get a more refined understanding of the evolution of RSF parameters, we estimated these parameters for each of the hundreds of stick-slip events that occurred on laboratory faults during an experiment. We used experiment data for which the gouge particles were removed from the laboratory faults at the beginning of each experiment; this procedure made possible to evaluate the influence of the gouge layer evolution on the variation of the RSF parameters. Since the amount of data was very large, we adopted a random forest (RF) machine learning approach for data analysis. The RF model was trained on simulated friction data and then applied to the experiment stick-slip event data to estimate the RSF parameters. To generate simulated friction data of stick-slip events, a one-degree-of-freedom spring-slider model governed by the RSF law was assumed. From plots of friction change as a function of displacement, some representative features were extracted to account for the RSF parameters and were used as input to the RF algorithm. Using the RF approach, we captured the variation of the RSF parameters a , |$b - a$| and |${D}_{\mathrm{c}}\ $| defined in the RSF law. The results show that during a first transient phase, the parameter a becomes smaller, while parameters |$b - a$| and |${D}_{\mathrm{c}}$| become larger, as the gouge layer becomes thicker. The variation of the RSF parameters becomes less pronounced during the following steady-state phase. These results suggest that the variation of RSF friction parameters may be related to the evolution of the gouge layer. [ABSTRACT FROM AUTHOR]

In the context of climate change, in-season and longer-term yield predictions are needed to anticipate local and regional food crises and propose adaptations to farmers' practices. Mechanistic models and machine learning are two modelling options to consider from this perspective. In this study, multiple regression (MR) and random forest (RF) models were calibrated for wheat yield prediction in Morocco, using data collected from 125 farmers' wheat fields. Additionally, MR and RF models were calibrated both with or without remotely sensed leaf area index (LAI), while considering all farmers' fields, or specifically to agroecological zoning in Morocco. The same farmers' fields were simulated using a mechanistic model (APSIM-wheat). We compared the predictive performances of the empirical models and APSIM-wheat. Results showed that both MR and RF showed rather good predictive quality (normalized root mean square errors (NRMSEs) below 35 %), but were always outperformed by the APSIM model. Both RF and MR selected remotely sensed LAI at heading, climate variables (maximal temperatures at emergence and tillering), and fertilization practices (amount of nitrogen applied at heading) as major yield predictors. Integration of remotely sensed LAI in the calibration process reduced NRMSE by 4.5 % and 1.8 % on average for MR and RF models, respectively. Calibration of region-specific models did not significantly improve the predictive. These findings lead to the conclusion that mechanistic models are better at capturing the impacts of in-season climate variability and would be preferred to support short-term tactical adjustments to farmers' practices, while machine learning models are easier to use in the perspective of mid-term regional prediction. [ABSTRACT FROM AUTHOR]

The underrepresentation of women in open-source software is frequently attributed to women's lack of innate aptitude compared to men: natural gender differences in technical ability (Trinkenreich et al. 2021). Approaching code as a form of communication, I conduct a novel empirical study of gender differences in Python programming on GitHub. Based on 1,728 open-source projects, I ask if there is a gender difference in the quality and style of Python code measured in adherence to PEP-8 guidelines. I found significant gender differences in structure and how Python files are organized. While there is gendered variation in programming style, there is no evidence of gender difference in code quality. Using a Random Forest model, I show that the gender of a programmer can be predicted from the style of their Python code. The study concludes that gender differences in Python code are a matter of style, not quality. [ABSTRACT FROM AUTHOR]

Background We aimed to develop a machine-learning model for predicting treatment response to radioiodine (131I) therapy and thyrotropin (TSH) suppression therapy in patients with differentiated thyroid cancer (DTC) but without structural disease, based on pre-treatment information. Patients and Methods Overall, 597 and 326 patients with DTC but without structural disease were randomly assigned to "training" cohorts for predicting treatment response to 131I therapy and TSH suppression therapy, respectively. Six supervised algorithms, including Logistic Regression, Support Vector Machine, Random Forest (RF), Neural Networks, Adaptive Boosting, and Gradient Boost, were used to predict effective response (ER) to 131I therapy and biochemical remission (BR) to TSH suppression therapy. Results Stimulated and suppressed thyroglobulin (Tg) and radioiodine uptake before the current course of 131I therapy were mostly attributed to ER to 131I therapy, while thyroid remnant available on the post-therapeutic whole-body scan at the last course of 131I therapy and TSH were greatly contributed to Tg decline under TSH suppression therapy. RF showed the best performance among all models. The accuracy and area under the receiver operating characteristic curve (AUC) for segregating ER from non-ER during 131I therapy with RF were 81.3% and 0.896, respectively. The accuracy and AUC for predicting BR to TSH suppression therapy with RF were 78.7% and 0.857, respectively. Conclusion This study demonstrates that machine learning models, especially the RF algorithm are useful tools that may predict treatment response to 131I therapy and TSH suppression therapy in DTC patients without structural disease based on pre-treatment routine clinical variables and biochemical markers. [ABSTRACT FROM AUTHOR]

We present a novel approach for the dimensionality reduction of galaxy images by leveraging a combination of variational auto-encoders (VAEs) and domain adaptation (DA). We demonstrate the effectiveness of this approach using a sample of low-redshift galaxies with detailed morphological type labels from the Galaxy Zoo Dark Energy Camera Legacy Survey (DECaLS) project. We show that 40-dimensional latent variables can effectively reproduce most morphological features in galaxy images. To further validate the effectiveness of our approach, we utilized a classical random forest classifier on the 40-dimensional latent variables to make detailed morphology feature classifications. This approach performs similar to a direct neural network application on galaxy images. We further enhance our model by tuning the VAE network via DA using galaxies in the overlapping footprint of DECaLS and Beijing-Arizona Sky Survey + Mayall z -band Legacy Survey, enabling the unbiased application of our model to galaxy images in both surveys. We observed that DA led to even better morphological feature extraction and classification performance. Overall, this combination of VAE and DA can be applied to achieve image dimensionality reduction, defect image identification, and morphology classification in large optical surveys. [ABSTRACT FROM AUTHOR]

In general relativity, approximations based on the spherical collapse model such as Press–Schechter theory and its extensions are able to predict the number of objects of a certain mass in a given volume. In this paper, we use a machine learning algorithm to test whether such approximations hold in screened modified gravity theories. To this end, we train random forest classifiers on data from N -body simulations to study the formation of structures in lambda cold dark matter (ΛCDM) as well as screened modified gravity theories, in particular f (R) and nDGP gravity. The models are taught to distinguish structure membership in the final conditions from spherical aggregations of density field behaviour in the initial conditions. We examine the differences between machine learning models that have learned structure formation from each gravity, as well as the model that has learned from ΛCDM. We also test the generalizability of the ΛCDM model on data from f (R) and nDGP gravities of varying strengths, and therefore the generalizability of extended Press–Schechter spherical collapse to these types of modified gravity. [ABSTRACT FROM AUTHOR]

We report on exceptionally bright bursts (>400 Jy ms) detected from the repeating fast radio burst source FRB 20220912A using the Nançay radio telescope (NRT), as part of the ECLAT (Extragalactic Coherent Light from Astrophysical Transients) monitoring campaign. These bursts exhibit extremely luminous, broad-band, short-duration structures (∼16  μ s), which we term 'microshots' and which can be especially well studied in the NRT data given the excellent signal-to-noise and dynamic range (32-bit samples). The estimated peak flux density of the brightest microshot is 450 Jy. We show that the microshots are clustered into dense 'forests' by modelling them as Weibull distributions and obtaining Weibull shape parameters of approximately 0.5. Our polarimetric analysis reveals that the bursts are nearly 100 per cent linearly polarized; have ≲10 per cent circular polarization fractions; a near-zero average rotation measure of 0.10(6) rad m−2; and varying polarization position angles over the burst duration. For one of the bursts, we analyse raw voltage data from simultaneous observations with the Westerbork RT-1 single 25-m dish. These data allow us to measure the scintillation bandwidth, 0.30(3) MHz, and to probe the bursts on (sub-)microsecond time-scales. Some important nuances related to dedispersion are also discussed. We propose that the emission mechanism for the broad-band microshots is potentially different from the emission mechanism of the broader burst components, which still show a residual drift of a few hundred MHz ms−1 after correcting for dispersion using the microshots. We discuss how the observed emission is phenomenologically analogous to different types of radio bursts from the Sun. [ABSTRACT FROM AUTHOR]

An equivalent–width-based classification may cause the erroneous judgement to the flat spectrum radio quasars (FSRQs) and BL Lacerate objects (BL Lac) due to the diluting the line features by dramatic variations in the jet continuum flux. To help address the issue, this work explores the possible intrinsic classification on the basis of a random forest supervised machine learning algorithm. In order to do so, we compile a sample of 1680 Fermi blazars that have both gamma-rays and radio-frequencies data available from the 4LAC-DR2 catalogue, which includes 1352 training and validation samples and 328 forecast samples. By studying the results for all of the different combinations of 23 characteristic parameters, we found that there are 178 optimal parameter combinations (OPCs) with the highest accuracy (≃98.89 per cent). Using the combined classification results from the nine combinations of these OPCs to the 328 forecast samples, we predict that there are 113 true BL Lacs (TBLs) and 157 false BL Lacs (FBLs) that are possible intrinsically FSRQs misclassified as BL Lacs. The FBLs show a clear separation from TBLs and FSRQs in the gamma-ray photon spectral index, Γph, and X -band radio flux, log FR , plot. Phenomenally, existence a BL Lac to FSRQ (B-to-F) transition zone is suggested, where the FBLs are in the stage of transition from BL Lacs to FSRQs. Comparing the LSP changing-look blazars (CLBs) reported in the literatures, the majority of LSP CLBs are located at the B-to-F zone. We argue that the FBLs located at B-to-F transition zone are the most likely candidates of CLBs. [ABSTRACT FROM AUTHOR]

Classifying active galactic nuclei (AGNs) is a challenge, especially for BL Lacertae objects (BLLs), which are identified by their weak emission line spectra. To address the problem of classification, we use data from the fourth Fermi Catalog, Data Release 3. Missing data hinder the use of machine learning to classify AGNs. A previous paper found that Multivariate Imputation by Chain Equations (MICE) imputation is useful for estimating missing values. Since many AGNs have missing redshift and the highest energy, we use data imputation with MICE and k -nearest neighbours (kNN) algorithm to fill in these missing variables. Then, we classify AGNs into the BLLs or the flat spectrum radio quasars (FSRQs) using the SuperLearner, an ensemble method that includes several classification algorithms like logistic regression, support vector classifiers, Random Forest, Ranger Random Forest, multivariate adaptive regression spline (MARS), Bayesian regression, and extreme gradient boosting. We find that a SuperLearner model using MARS regression and Random Forest algorithms is 91.1 per cent accurate for kNN-imputed data and 91.2 per cent for MICE-imputed data. Furthermore, the kNN-imputed SuperLearner model predicts that 892 of the 1519 unclassified blazars are BLLs and 627 are FSRQs, while the MICE-imputed SuperLearner model predicts 890 BLLs and 629 FSRQs in the unclassified set. Thus, we can conclude that both imputation methods work efficiently and with high accuracy and that our methodology ushers the way for using SuperLearner as a novel classification method in the AGN community and, in general, in the astrophysics community. [ABSTRACT FROM AUTHOR]

Automatic detection of seismic events in ocean bottom seismometer (OBS) data is difficult due to elevated levels of noise compared to the recordings from land. Popular deep-learning approaches that work well with earthquakes recorded on land perform poorly in a marine setting. Their adaptation to OBS data requires catalogues containing hundreds of thousands of labelled event examples that currently do not exist, especially for signals different than earthquakes. Therefore, the usual routine involves standard amplitude-based detection methods and manual processing to obtain events of interest. We present here the first attempt to utilize a Random Forest supervised machine learning classifier on marine seismological data to automate catalogue screening and event recognition among different signals [i.e. earthquakes, short duration events (SDE) and marine noise sources]. The detection approach uses the short-term average/long-term average method, enhanced by a kurtosis-based picker for a more precise recognition of the onset of events. The subsequent machine learning method uses a previously published set of signal features (waveform-, frequency- and spectrum-based), applied successfully in recognition of different classes of events in land seismological data. Our workflow uses a small subset of manually selected signals for the initial training procedure and we then iteratively evaluate and refine the model using subsequent OBS stations within one single deployment in the eastern Fram Strait, between Greenland and Svalbard. We find that the used set of features is well suited for the discrimination of different classes of events during the training step. During the manual verification of the automatic detection results, we find that the produced catalogue of earthquakes contains a large number of noise examples, but almost all events of interest are properly captured. By providing increasingly larger sets of noise examples we see an improvement in the quality of the obtained catalogues. Our final model reaches an average accuracy of 87 per cent in recognition between the classes, comparable to classification results for data from land. We find that, from the used set of features, the most important in separating the different classes of events are related to the kurtosis of the envelope of the signal in different frequencies, the frequency with the highest energy and overall signal duration. We illustrate the implementation of the approach by using the temporal and spatial distribution of SDEs as a case study. We used recordings from six OBSs deployed between 2019 and 2020 off the west-Svalbard coast to investigate the potential link of SDEs to fluid dynamics and discuss the robustness of the approach by analysing SDE intensity, periodicity and distance to seepage sites in relation to other published studies on SDEs. [ABSTRACT FROM AUTHOR]

Machine learning feature importance calculations are used to determine the molecular substructures that are responsible for mid- and far-infrared (IR) emission features of neutral polycyclic aromatic hydrocarbons (PAHs). Using the extended-connectivity fingerprint as a descriptor of chemical structure, a random forest model is trained on the spectra of 14 124 PAHs to evaluate the importance of 10 632 molecular fragments for each band within the range of 2.761 to |$1172.745\, \mu$| m. The accuracy of the results is confirmed by comparing them with previously studied unidentified infrared emission (UIE) bands. The results are summarized in two tables available as Supplementary Data , which can be used as a reference for assessing possible UIE carriers. We demonstrate that the tables can be used to explore the relation between the PAH structure and the spectra by discussing about the IR features of nitrogen-containing PAHs and superhydrogenated PAHs. [ABSTRACT FROM AUTHOR]

Background Although lung cancer screening with low-dose computed tomography is rolling out in many areas of the world, differentiating indeterminate pulmonary nodules remains a major challenge. We conducted one of the first systematic investigations of circulating protein markers to differentiate malignant from benign screen-detected pulmonary nodules. Methods Based on 4 international low-dose computed tomography screening studies, we assayed 1078 protein markers using prediagnostic blood samples from 1253 participants based on a nested case-control design. Protein markers were measured using proximity extension assays, and data were analyzed using multivariable logistic regression, random forest, and penalized regressions. Protein burden scores (PBSs) for overall nodule malignancy and imminent tumors were estimated. Results We identified 36 potentially informative circulating protein markers differentiating malignant from benign nodules, representing a tightly connected biological network. Ten markers were found to be particularly relevant for imminent lung cancer diagnoses within 1 year. Increases in PBSs for overall nodule malignancy and imminent tumors by 1 standard deviation were associated with odds ratios of 2.29 (95% confidence interval: 1.95 to 2.72) and 2.81 (95% confidence interval: 2.27 to 3.54) for nodule malignancy overall and within 1 year of diagnosis, respectively. Both PBSs for overall nodule malignancy and for imminent tumors were substantially higher for those with malignant nodules than for those with benign nodules, even when limited to Lung Computed Tomography Screening Reporting and Data System (LungRADS) category 4 (P  < .001). Conclusions Circulating protein markers can help differentiate malignant from benign pulmonary nodules. Validation with an independent computed tomographic screening study will be required before clinical implementation. [ABSTRACT FROM AUTHOR]

The full-shape correlations of the Lyman alpha (Ly α) forest contain a wealth of cosmological information through the Alcock–Paczyński effect. However, these measurements are challenging to model without robustly testing and verifying the theoretical framework used for analysing them. Here, we leverage the accuracy and volume of the N -body simulation suite AbacusSummit to generate high-resolution Ly α skewers and quasi-stellar object (QSO) catalogues. One of the main goals of our mocks is to aid in the full-shape Ly α analysis planned by the Dark Energy Spectroscopic Instrument (DESI) team. We provide optical depth skewers for six of the fiducial cosmology base-resolution simulations (⁠|$L_{\rm box} = 2\, h^{-1}\, {\rm Gpc}$|⁠ , N  = 69123) at z  = 2.5. We adopt a simple recipe based on the Fluctuating Gunn–Peterson Approximation (FGPA) for constructing these skewers from the matter density in an N -body simulation and calibrate it against the 1D and 3D Ly α power spectra extracted from the hydrodynamical simulation IllustrisTNG (TNG; |$L_{\rm box} = 205\, h^{-1}\, {\rm Mpc}$|⁠ , N  = 25003). As an important application, we study the non-linear broadening of the baryon acoustic oscillation (BAO) peak and show the cross-correlation between DESI-like QSOs and our Ly α forest skewers. We find differences on small scales between the Kaiser approximation prediction and our mock measurements of the Ly α × QSO cross-correlation, which would be important to account for in upcoming analyses. The AbacusSummit Ly α forest mocks open up the possibility for improved modelling of cross-correlations between Ly α and cosmic microwave background (CMB) lensing and Ly α and QSOs, and for forecasts of the 3-point Ly α correlation function. Our catalogues and skewers are publicly available on Globus via the National Energy Research Scientific Computing Center (NERSC) (full link under the section 'Data Availability'). [ABSTRACT FROM AUTHOR]

Most of water reservoirs are underground and therefore challenging to monitor. This is particularly the case of karst aquifers which knowledge is mostly based on sparse spatial and temporal observations. In this study, we propose a new approach, based on a supervised machine learning algorithm, the Random Forests, and continuous seismic noise records, that allows the prediction of the underground river water height. The study site is a karst aquifer in the Jura Mountains (France). An underground river is accessible through an artificial shaft and is instrumented by a hydrological probe. The seismic noise generated by the river is recorded by two broadband seismometers, located underground (20 m depth) and at the surface. The algorithm succeeds in predicting water height thanks to signal energy features. Even weak river-induced noise such as recorded at the surface can be detected and used by the algorithm. Its efficiency, expressed by the Nash–Sutcliffe criterion, is above 95 per cent and 53 per cent for data from the underground and surface seismic stations, respectively. [ABSTRACT FROM AUTHOR]

We develop a random forest regressor (RFR) machine learning model to trace the coronal evolution in two highly variable active galactic nuclei (AGNs) IRAS 13224-3809 and 1H 0707-495 observed with XMM–Newton , by probing the X-ray reverberation features imprinted on their power spectral density (PSD) profiles. Simulated PSDs in the form of a power law, with a similar frequency range and bins to the observed data, are produced. Then, they are convolved with relativistic disc-response functions from a lamp-post source before being used to train and test the model to predict the coronal height. We remove some bins that are dominated by Poisson noise and find that the model can tolerate the frequency-bin removal up to ∼10 bins to maintain a prediction accuracy of R 2 > 0.9. The black hole mass and inclination should be fixed so that the accuracy in predicting the source height is still >0.9. The accuracy also increases with the reflection fraction. The corona heights for both AGN are then predicted using the RFR model developed from the simulated PSDs, whose frequency range and bins are specifically adjusted to match those from each individual observation. The model suggests that their corona varies between ∼5–18  r g, with R 2 > 0.9 for all observations. Such high accuracy can still be obtained if the difference between the true mass and the trained value is |${\lesssim}10\ \hbox{per cent}$|⁠. Finally, the model supports the height-changing corona under the light-bending scenario, where the height is correlated to source luminosity in both IRAS 13224-3809 and 1H 0707-495. [ABSTRACT FROM AUTHOR]

Motivation Random forest is a popular machine learning approach for the analysis of high-dimensional data because it is flexible and provides variable importance measures for the selection of relevant features. However, the complex relationships between the features are usually not considered for the selection and thus also neglected for the characterization of the analysed samples. Results Here we propose two novel approaches that focus on the mutual impact of features in random forests. Mutual forest impact (MFI) is a relation parameter that evaluates the mutual association of the features to the outcome and, hence, goes beyond the analysis of correlation coefficients. Mutual impurity reduction (MIR) is an importance measure that combines this relation parameter with the importance of the individual features. MIR and MFI are implemented together with testing procedures that generate P -values for the selection of related and important features. Applications to one experimental and various simulated datasets and the comparison to other methods for feature selection and relation analysis show that MFI and MIR are very promising to shed light on the complex relationships between features and outcome. In addition, they are not affected by common biases, e.g. that features with many possible splits or high minor allele frequencies are preferred. Availability and implementation The approaches are implemented in Version 0.3.3 of the R package RFSurrogates that is available at github.com/AGSeifert/RFSurrogates and the data are available at doi.org/10.25592/uhhfdm.12620. [ABSTRACT FROM AUTHOR]

Objectives To identify highly ranked features related to clinicians' diagnosis of clinically relevant knee OA. Methods General practitioners (GPs) and secondary care physicians (SPs) were recruited to evaluate 5–10 years follow-up clinical and radiographic data of knees from the CHECK cohort for the presence of clinically relevant OA. GPs and SPs were gathered in pairs; each pair consisted of one GP and one SP, and the paired clinicians independently evaluated the same subset of knees. A diagnosis was made for each knee by the GP and SP before and after viewing radiographic data. Nested 5-fold cross-validation enhanced random forest models were built to identify the top 10 features related to the diagnosis. Results Seventeen clinician pairs evaluated 1106 knees with 139 clinical and 36 radiographic features. GPs diagnosed clinically relevant OA in 42% and 43% knees, before and after viewing radiographic data, respectively. SPs diagnosed in 43% and 51% knees, respectively. Models containing top 10 features had good performance for explaining clinicians' diagnosis with area under the curve ranging from 0.76–0.83. Before viewing radiographic data, quantitative symptomatic features (i.e. WOMAC scores) were the most important ones related to the diagnosis of both GPs and SPs; after viewing radiographic data, radiographic features appeared in the top lists for both, but seemed to be more important for SPs than GPs. Conclusions Random forest models presented good performance in explaining clinicians' diagnosis, which helped to reveal typical features of patients recognized as clinically relevant knee OA by clinicians from two different care settings. [ABSTRACT FROM AUTHOR]

Motivation Tuberculosis (TB) is caused by members of the Mycobacterium tuberculosis complex (MTBC), which has a strain- or lineage-based clonal population structure. The evolution of drug-resistance in the MTBC poses a threat to successful treatment and eradication of TB. Machine learning approaches are being increasingly adopted to predict drug-resistance and characterize underlying mutations from whole genome sequences. However, such approaches may not generalize well in clinical practice due to confounding from the population structure of the MTBC. Results To investigate how population structure affects machine learning prediction, we compared three different approaches to reduce lineage dependency in random forest (RF) models, including stratification, feature selection, and feature weighted models. All RF models achieved moderate-high performance (area under the ROC curve range: 0.60–0.98). First-line drugs had higher performance than second-line drugs, but it varied depending on the lineages in the training dataset. Lineage-specific models generally had higher sensitivity than global models which may be underpinned by strain-specific drug-resistance mutations or sampling effects. The application of feature weights and feature selection approaches reduced lineage dependency in the model and had comparable performance to unweighted RF models. Availability and implementation https://github.com/NinaMercedes/RF_lineages. [ABSTRACT FROM AUTHOR]

Studies have confirmed that the occurrence of many complex diseases in the human body is closely related to the microbial community, and microbes can affect tumorigenesis and metastasis by regulating the tumor microenvironment. However, there are still large gaps in the clinical observation of the microbiota in disease. Although biological experiments are accurate in identifying disease-associated microbes, they are also time-consuming and expensive. The computational models for effective identification of diseases related microbes can shorten this process, and reduce capital and time costs. Based on this, in the paper, a model named DSAE_RF is presented to predict latent microbe–disease associations by combining multi-source features and deep learning. DSAE_RF calculates four similarities between microbes and diseases, which are then used as feature vectors for the disease-microbe pairs. Later, reliable negative samples are screened by k -means clustering, and a deep sparse autoencoder neural network is further used to extract effective features of the disease-microbe pairs. In this foundation, a random forest classifier is presented to predict the associations between microbes and diseases. To assess the performance of the model in this paper, 10-fold cross-validation is implemented on the same dataset. As a result, the AUC and AUPR of the model are 0.9448 and 0.9431, respectively. Furthermore, we also conduct a variety of experiments, including comparison of negative sample selection methods, comparison with different models and classifiers, Kolmogorov–Smirnov test and t -test, ablation experiments, robustness analysis, and case studies on Covid-19 and colorectal cancer. The results fully demonstrate the reliability and availability of our model. [ABSTRACT FROM AUTHOR]

Background Kidneys donated after circulatory death suffer a period of functional warm ischaemia before death, which may lead to early ischaemic injury. Effects of haemodynamic trajectories during the agonal phase on delayed graft function (DGF) is unknown. We aimed to predict the risk of DGF using patterns of trajectories of systolic blood pressure (SBP) declines in Maastricht category 3 kidney donors. Methods We conducted a cohort study of all kidney transplant recipients in Australia who received kidneys from donation after circulatory death donors, divided into a derivation cohort (transplants between 9 April 2014 and 2 January 2018 [462 donors]) and a validation cohort (transplants between 6 January 2018 and 24 December 2019 [324 donors]). Patterns of SBP decline using latent class models were evaluated against the odds of DGF using a two-stage linear mixed effects model. Results In the derivation cohort, 462 donors were included in the latent class analyses and 379 donors in the mixed effects model. Of the 696 eligible transplant recipients, 380 (54.6%) experienced DGF. Ten different trajectories, with distinct patterns of SBP decline were identified. Compared with recipients from donors with the slowest decline in SBP after withdrawal of cardiorespiratory support, the adjusted odds ratio (aOR) for DGF was 5.5 [95% confidence interval (CI) 1.38–28.0] for recipients from donors with a steeper decline and lowest SBP [mean 49.5 mmHg (standard deviation 12.5)] at the time of withdrawal. For every 1 mmHg/min reduction in the rate of decline of SBP, the respective aORs for DGF were 0.95 (95% CI 0.91–0.99) and 0.98 (95% CI 0.93–1.0) in the random forest and least absolute shrinkage and selection operator models. In the validation cohort, the respective aORs were 0.95 (95% CI 0.91–1.0) and 0.99 (95% CI 0.94–1.0). Conclusion Trajectories of SBP decline and their determinants are predictive of DGF. These results support a trajectory-based assessment of haemodynamic changes in donors after circulatory death during the agonal phase for donor suitability and post-transplant outcomes. [ABSTRACT FROM AUTHOR]