Machine Learning-based Identification Research Articles

Machine learning-based identification of proteomic markers in colorectal cancer using UK Biobank data.

Colorectal cancer is one of the leading causes of cancer-related mortality in the world. Incidence and mortality are predicted to rise globally during the next several decades. When detected early, colorectal cancer is treatable with surgery and medications. This leads to the requirement for prognostic and diagnostic biomarker development. Our study integrates machine learning models and protein network analysis to identify protein biomarkers for colorectal cancer. Our methodology leverages an extensive collection of proteome profiles from both healthy and colorectal cancer individuals. To identify a potential biomarker with high predictive ability, we used three machine learning models. To enhance the interpretability of our models, we quantify each protein's contribution to the model's predictions using SHapley Additive exPlanations values. Three classifiers-LASSO, XGBoost, and LightGBM were evaluated for predictive performance along with hyperparameter tuning of each model using grid search, with LASSO achieving the highest AUC of 75% in the UK Biobank dataset and the AUCs for LightGBM and XGBoost are 69.61% and 71.42%, respectively. Using SHapley Additive exPlanations values, TFF3, LCN2, and CEACAM5 were found to be key biomarkers associated with cell adhesion and inflammation. Protein quantitative trait loci analyze studies provided further evidence for the involvement of TFF1, CEACAM5, and SELE in colorectal cancer, with possible connections to the PI3K/Akt and MAPK signaling pathways. By offering insights into colorectal cancer diagnostics and targeted therapeutics, our findings set the stage for further biomarker validation.

Machine learning-based identification and validation of amino acid metabolism related genes as novel biomarkers in chronic kidney disease

Machine learning-based identification and validation of amino acid metabolism related genes as novel biomarkers in chronic kidney disease

Machine learning-based identification of animal feeding operations in the United States on a parcel-scale.

The increasing global demand for meat and dairy products, fueled by rapid industrialization, has led to the expansion of Animal Feeding Operations (AFOs) in the United States (US). These operations, often found in clusters, generate large amounts of manure, posing a considerable risk to water quality due to the concentrated waste streams they produce. Accurately mapping AFOs is essential for effective environmental and disease management, yet many facilities remain undocumented due to variations in federal and state regulations. Current techniques for mapping AFOs in the US rely on a mix of manual digitization, aerial imaging, and image processing. By applying a machine learning-based random forest (RF) classification method to a socio-environmental dataset that excluded aerial images in this work, we overcame some of the limitations associated with aerial image-based approaches, enhancing mapping accuracy to 87%. We used publicly available environmental, nutrient-focused, and socioeconomic data downscaled to the parcel level, which more accurately reflects farm boundaries and operations than previous methods. Our study incorporates 58 variables, with canopy cover, surrounding vegetation, day and nighttime land surface temperatures, and phosphorus from animals identified as key predictors of AFO presence. The relevance of these variables varies across states, influenced by whether the dominant land covers are human-induced, like croplands, or natural, such as savannas and grasslands. Thus, our public-data based approach, easily replicable, not only improves the precision of AFO detection, but also facilitates the monitoring of nutrient flows at the parcel level-critical for nutrient budgeting and recovery, water quality management, and disease risk assessment and tracing.

Machine learning-based identification of histone deacetylase-associated prognostic factors and prognostic modeling for low-grade glioma.

Low-grade glioma (LGG) is a slow-growing but invasive tumor that affects brain function. Histone deacetylases (HDACs) play a critical role in gene regulation and tumor progression. This study aims to develop a prognostic model based on HDAC-related genes to aid in risk stratification and predict therapeutic responses. Expression data from The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) were analyzed to identify an optimal HDAC-related risk signature from 73 genes using 10 machine learning algorithms. Patients were stratified into high- and low-risk groups based on the median risk score. Prognostic accuracy was evaluated using Kaplan-Meier survival analysis and receiver operating characteristic (ROC) curves. Functional enrichment analyses, including Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA), were performed to explore pathways linked to the gene signature. Immune infiltration and tumor microenvironment characteristics were assessed using Single Sample Gene Set Enrichment Analysis (ssGSEA) and ESTIMATE algorithm. SubMap was applied to predict responsiveness to immune checkpoint inhibitors, and chemotherapeutic sensitivity was analyzed via the Genomics of Drug Sensitivity in Cancer (GDSC) database. A prognostic model consisting of four HDAC-related genes-SP140, BAZ1B, SP100, and SIRT1-was identified. This signature displayed strong prognostic accuracy, achieving a C-index of 0.945. Individuals with LGG were systematically divided into high-risk and low-risk cohorts based on the median risk value, enabling more precise risk stratification. The survival prognosis was significantly worse in the high-risk cohort compared to the low-risk group, highlighting distinct survival trajectories. Notably, the two cohorts exhibited marked shifts in immune checkpoint gene transcriptional profiles and immune cell infiltration maps, underscoring fundamental biological differences that contribute to these differing prognoses. We developed an HDAC-related four-gene prognostic model that correlates with survival, immune landscape, and therapeutic response in LGG patients. This model may guide personalized treatment strategies and improve prognostic accuracy, warranting further validation in clinical settings.

Machine Learning-Based Identification of High-Risk Patterns in Atrial Fibrillation Ablation Outcomes.

Atrial fibrillation (AF) is one of the most common types of cardiac arrhythmias, often leading to serious health issues such as stroke, heart failure, and higher mortality rates. Its global impact is rising due to aging populations and growing comorbidities, creating an urgent need for more effective treatment methods. AF ablation, a key treatment option, has success rates that vary widely among patients. Conventional predictors of ablation outcomes, which primarily rely on sociodemographic and clinical factors, fall short of capturing the heterogeneity within patient populations, highlighting the potential for data-driven methods to provide deeper insights into procedural success. To uncover meaningful patient subgroups based on AF ablation outcomes and identify diagnostic codes associated with failure. Machine learning clustering with must-link and cannot-link constraints was applied to electronic health records to discover meaningful clusters, revealing patient-specific factors influencing procedural success or failure. Statistical analyses, including chi-square tests, were used to identify diagnostic codes significantly associated with ablation failure. Out of the 145 diagnostic codes examined, thirteen significant codes were identified and categorized into four primary risk groups, ranked by their impact on procedural outcomes: (1) direct contributors affecting cardiovascular health, (2) indirect factors that contribute to systemic stress, (3) complications related to anticoagulation and hemorrhagic risks that can impact bleeding management, and (4) broader health indicators reflecting a general health burden that reduce patients resilience to procedural stress. This study shows the importance of cardiovascular and non-cardiovascular factors in AF ablation outcomes, emphasizing the need for a more comprehensive pre-procedural evaluation. It also contributes to the application of machine learning in personalized risk assessment for AF and advancing individualized care strategies that may improve ablation success.

Machine learning-based identification of critical factors for cadmium accumulation in rice grains.

The aggregation of Cadmium (Cd) in rice grains is a significant threat to human healthy. The complexity of the soil-rice system, with its numerous influencing parameters, highlights the need to identify the crucial factors responsible for Cd aggregation. This study uses machine learning (ML) modeling to predict Cd aggregation in rice grains and identify the influencing factors. Data from 474 data points from 77 published works were analyzed, and eight ML models were established using different algorithms. The input variables were total soil Cd concentration (TS Cd) and extractable Cd concentration (Ex-Cd), while rice Cd concentration (Cdrice) was the output variable. Among the models, the Extremely Randomized Trees (ERT) model performed the best (TS Cd: R2 = 0.825; Ex-Cd: R2 = 0.792), followed by Random Forest (TS Cd: R2 = 0.721; Ex-Cd: R2 = 0.719). The ERT feature importance ranking analysis revealed that the essential factors responsible for Cd aggregation are cation exchange capacity (CEC), TS Cd, Water Management Model (WMM), and pH for total soil Cd as input variables. For extractable Cd as an input variable, the vital factors are CEC, Ex-Cd, pH, and WMM. The study highlights the importance of the Water Management Model and its impact on Cd concentration in rice grains, which has been overlooked in previous research.Please check and confirm that the authors and their respective affiliations have been correctly identified and amend if necessary.The authors and their respective affiliations are correct.Author details: Kindly check and confirm whether the corresponding author is correctly identified.It is correct.

Machine learning-based identification of general transcriptional predictors for plant disease.

This study investigated the generalizability of Arabidopsis thaliana immune responses across diverse pathogens, including Botrytis cinerea, Sclerotinia sclerotiorum, and Pseudomonas syringae, using a data-driven, machine learning approach. Machine learning models were trained to predict disease development from early transcriptional responses. Feature selection techniques based on network science and topology were used to train models employing only a fraction of the transcriptome. Machine learning models trained on one pathosystem where then validated by predicting disease development in new pathosystems. The identified feature selection gene sets were enriched for pathways related to biotic, abiotic, and stress responses, though the specific genes involved differed between feature sets. This suggests common immune responses to diverse pathogens that operate via different gene sets. The study demonstrates that machine learning can uncover both established and novel components of the plant's immune response, offering insights into disease resistance mechanisms. These predictive models highlight the potential to advance our understanding of multigenic outcomes in plant immunity and can be further refined for applications in disease prediction.

Topological embedding and directional feature importance in ensemble classifiers for multi-class classification

Cancer is the second leading cause of disease-related death worldwide, and machine learning-based identification of novel biomarkers is crucial for improving early detection and treatment of various cancers. A key challenge in applying machine learning to high-dimensional data is deriving important features in an interpretable manner to provide meaningful insights into the underlying biological mechanismsWe developed a class-based directional feature importance (CLIFI) metric for decision tree methods and demonstrated its use for The Cancer Genome Atlas proteomics data. The CLIFI metric was incorporated into four algorithms, Random Forest (RF), LAtent VAriable Stochastic Ensemble of Trees (LAVASET), and Gradient Boosted Decision Trees (GBDTs), and a new extension incorporating the LAVA step into GBDTs (LAVABOOST). Both LAVA methods incorporate topological information from protein interactions into the decision function.The different models' performance in classifying 28 cancers resulted in F1-scores of 92.6% (RF), 92.0% (LAVASET), 89.3% (LAVABOOST) and 85.7% (GBDT), with no method outperforming all others for individual cancer type prediction. The CLIFI metric enables visualisation of the model's decision-making functions. The resulting CLIFI value distributions indicated heterogeneity in the expression of several proteins (MYH11, ERα, BCL2) across different cancer types (including brain glioma, breast, kidney, thyroid and prostate cancer) aligning with the original raw expression data.In conclusion, we have developed an integrated, directional feature importance metric for multi-class decision tree-based classification models that facilitates interpretable feature importance assessment. The CLIFI metric can be combined with incorporating topological information into the decision functions of models to introduce inductive bias, enhancing interpretability.

Machine learning-based identification of geomorphological units in Quintero Bay (32°S) and its implications for the search for early drowned archaeological sites on the western coast of South America

High-resolution predictive modeling of submerged landscapes has successfully allowed the detection of early archaeological sites that are presently underwater. These models have traditionally relied on geophysical techniques, which can be both time-consuming and expensive, especially for extensive survey areas. In contrast, geomorphological mapping using Machine Learning (ML) techniques has emerged as a rapid and accessible alternative with numerous advantages over conventional methods. In this work, we employ ML algorithms (Random Forest, Support Vector Machine, Partial Least Squares, and Principal Component Analysis) trained on land to analyze the seabed of Quintero Bay to identify relic landforms that characterize the paleolandscape within which the submerged early site GNLQ1 formed. The methodology also included a multicriteria analysis that integrated geological (geomorphological, tectonic, eustatic) and archaeological (attributes of non-submerged records in the region) approaches to delineate potential areas of archaeological interest. The findings of this work can guide and enhance future archaeological research. The results underscore the importance of possessing a comprehensive understanding of the study area and its associated variables to the successful application of ML techniques. This also applies to modeling drowned paleolandscapes. Nevertheless, despite these challenges, ML-based modeling of drowned paleolandscapes can provide an overview of the distribution of geoforms comprising the paleolandscape, which in turn can help identify future geophysical survey areas to focus on in the search for archaeological evidence, thereby improving our understanding of the relationship between early human groups and these landscapes.

Machine learning-based identification and validation of immune-related biomarkers for early diagnosis and targeted therapy in diabetic retinopathy

The early diagnosis of diabetic retinopathy (DR) is challenging, highlighting the urgent need to identify new biomarkers. Immune responses play a crucial role in DR, yet there are currently no reports of machine learning (ML) algorithms being utilized for the development of immune-related molecular markers in DR. Based on the datasets GSE102485 and GSE160306, differentially expressed genes (DEGs) were screened using Weighted Gene Co-expression Network Analysis (WGCNA). Five ML algorithms including Bayesian, Learning Vector Quantization (LVQ), Wrapper (Boruta), Random Forest (RF), and Logistic Regression were employed to select immune-related genes associated with DR (DR.Sig). Seven ML algorithms including Naive Bayes (NB), RF, Support Vector Machine (SVM), AdaBoost Classification Trees (AdaBoost), Boosted Logistic Regressions (LogitBoost), K-Nearest Neighbors (KNN), and Cancerclass were utilized to construct a predictive model for DR. The relationship between DR.Sig genes and immune cells was analyzed using single-sample Gene Set Enrichment Analysis (ssGSEA). Additionally, drug sensitivity prediction of DR.Sig genes and molecular docking were performed. Through the utilization of 5 ML algorithms, 6 immune-related biomarkers closely related to the occurrence of DR were identified, including FCGR2B, CSRP1, EDNRA, SDC2, TEK, and CIITA. The DR predictive model constructed based on these 6 DR.Sig genes using the Cancerclass algorithm demonstrated superior predictive performance compared to 4 previously published DR-related biomarkers. In vivo and in vitro experiments also provided strong validation of the expression of the 6 genes in DR. Positive correlations were observed between these genes and 22 types of immune cells. Molecular docking results revealed that CSRP1, EDNRA, and TEK exhibited the highest affinities with the small molecule compounds etoposide, FR-139317, and camptothecin, respectively. The models constructed based on various ML algorithms can effectively predict the occurrence of DR events and hold potential for targeted drug therapies, providing a basis for the early diagnosis and targeted treatment of DR.

Machine learning-based identification and validation of aging-related genes in cardiomyocytes from patients with atrial fibrillation.

Aging is a key risk factor for atrial fibrillation (AF), a prevalent cardiac disorder among the elderly. This study aims to elucidate the genetic underpinnings of AF in the context of aging. We analyzed 12,403 genes from the GSE2240 database and 279 age-related genes from the CellAge database. Machine learning algorithms, including support vector machines and random forests, were employed to identify genes significantly associated with AF. Among the genes studied, 76 were found to be potential candidates in the development of AF. Notably, four genes - PTTG1, AR, RAD21, and YAP1 - stood out with a Receiver Operating Characteristic Area Under the Curve (ROC AUC) of 0.9, signifying high predictive power. Logistic regression, validated through 10-fold cross-validation and Bootstrap resampling, was determined as the most suitable model for internal validation. The discovery of these four genes could improve diagnostic accuracy for AF in the aged population. Additionally, our drug prediction model indicates that bisphenol A and cisplatin, among other substances, could be promising in treating age-associated AF, offering potential pathways for clinical intervention.

Machine learning-based identification of an immunotherapy-related signature to enhance outcomes and immunotherapy responses in melanoma.

Immunotherapy has revolutionized skin cutaneous melanoma treatment, but response variability due to tumor heterogeneity necessitates robust biomarkers for predicting immunotherapy response. We used weighted gene co-expression network analysis (WGCNA), consensus clustering, and 10 machine learning algorithms to develop the immunotherapy-related gene model (ITRGM) signature. Multi-omics analyses included bulk and single-cell RNA sequencing of melanoma patients, mouse bulk RNA sequencing, and pathology sections of melanoma patients. We identified 66 consensus immunotherapy prognostic genes (CITPGs) using WGCNA and differentially expressed genes (DEGs) from two melanoma cohorts. The CITPG-high group showed better prognosis and enriched immune activities. DEGs between CITPG-high and CITPG-low groups in the TCGA-SKCM cohort were analyzed in three additional melanoma cohorts using univariate Cox regression, resulting in 44 consensus genes. Using 101 machine learning algorithm combinations, we constructed the ITRGM signature based on seven model genes. The ITRGM outperformed 37 published signatures in predicting immunotherapy prognosis across the training cohort, three testing cohorts, and a meta-cohort. It effectively stratified patients into high-risk or low-risk groups for immunotherapy response. The low-risk group, with high levels of model genes, correlated with increased immune characteristics such as tumor mutation burden and immune cell infiltration, indicating immune-hot tumors with a better prognosis. The ITRGM's relationship with the tumor immune microenvironment was further validated in our experiments using pathology sections with GBP5, an important model gene, and CD8 IHC analysis. The ITRGM also predicted better immunotherapy response in eight cohorts, including urothelial carcinoma and stomach adenocarcinoma, indicating broad applicability. The ITRGM signature is a stable and robust predictor for stratifying melanoma patients into 'immune-hot' and 'immune-cold' tumors, enhancing prognosis and response to immunotherapy.

Machine learning-based identification of the strongest predictive features of scoring penalty kick in football

In football, the penalty is the situation that has one of the highest chances of scoring a goal. However, the success of a penalty kick highly depends on many kinds of attributes, including the penalty-takers’ abilities, the amount of fan pressure, the minute of the match, and the current score. In this paper, 16 features extracted from penalty kick positions, penalty-takers’ information, and match-day preferences and machine learning used to predict penalty kick result. Moreover, we revealed the most important feature combination that significantly affected the success of a penalty kick. The proposed method was trained with 120 and tested with 50 penalty kicks from the Turkish Super League in terms of classification accuracy and polygon area metric. We concluded that the result of a penalty kick can be predicted with a mean classification accuracy and mean polygon area metric rates of 79.80% and 0.60 using the k-nearest neighbor classifier.

Machine Learning-Based Identification of Risk Factors of Keratoconus Progression Using Raw Corneal Tomography Data.

The purpose of this study was to identify early indicators of keratoconus progression in Pentacam data using machine learning (ML) techniques. A retrospective Pentacam tabular data set was created by retrieving 11,760 tomography tests performed in patients with keratoconus. Data for eyes labeled unstable based on their referral for cross-linking were differentiated from data for eyes labeled stable and not referred for follow-up procedures. A boosted decision tree was trained on the final data set using a cross-validation method. The final labeled data set included 1218 tomography tests. Training a ML model on a single test for each eye did not accurately predict disease progression, as indicated by the mean receiver-operating characteristic area under the curve of 0.59 ± 0.1, with precision of 0.27, recall of 0.3, and F1 score of 0.28. Training on serial tests for each eye included 819 tomography scans and yielded good prognostic abilities: a receiver-operating characteristic area under the curve of 0.75 ± 0.07, precision of 0.32, recall of 0.67, and F1 score of 0.43. In addition, 4 of the 55 Pentacam raw data parameters predominantly used the algorithm decision: age, central keratoconus index, Rs B, and D10 mm pachy. This study revealed specific dominant parameters attributing to the classification of stability, which are not routinely assessed in determining progression in common practice. Using ML techniques, keratoconus deterioration was evaluated algorithmically with training on multiple tests, yet was not predicted by a single tomography test. Hence, our study highlights novel factors to the current consideration of cross-linking referral and may serve as a supportive tool for clinicians.

Machine learning-based identification of biomarkers and drugs in immunologically cold and hot pancreatic adenocarcinomas

BackgroundPancreatic adenocarcinomas (PAADs) often exhibit a “cold” or immunosuppressive tumor milieu, which is associated with resistance to immune checkpoint blockade therapy; however, the underlying mechanisms are incompletely understood. Here, we aimed to improve our understanding of the molecular mechanisms occurring in the tumor microenvironment and to identify biomarkers, therapeutic targets, and potential drugs to improve PAAD treatment.MethodsPatients were categorized according to immunologically hot or cold PAAD subtypes with distinct disease outcomes. Cox regression and weighted correlation network analysis were performed to construct a novel gene signature, referred to as ‘Downregulated in hot tumors, Prognostic, and Immune-Related Genes’ (DPIRGs), which was used to develop prognostic models for PAAD via machine learning (ML). The role of DPIRGs in PAAD was comprehensively analyzed, and biomarker genes able to distinguish PAAD immune subtypes and predict prognosis were identified by ML. The expression of biomarkers was verified using public single-cell transcriptomic and proteomic resources. Drug candidates for turning cold tumors hot and corresponding target proteins were identified via molecular docking studies.ResultsUsing the DPIRG signature as input data, a combination of survival random forest and partial least squares regression Cox was selected from 137 ML combinations to construct an optimized PAAD prognostic model. The effects and molecular mechanisms of DPIRGs were investigated by analysis of genetic/epigenetic alterations, immune infiltration, pathway enrichment, and miRNA regulation. Biomarkers and potential therapeutic targets, including PLEC, TRPV1, and ITGB4, among others, were identified, and the cell type-specific expression of the biomarkers was validated. Drug candidates, including thalidomide, SB-431542, and bleomycin A2, were identified based on their ability to modulate DPIRG expression favorably.ConclusionsBy combining multiple ML algorithms, we developed a novel prognostic model with excellent performance in PAAD cohorts. ML also proved to be powerful for identifying biomarkers and potential targets for improved PAAD patient stratification and immunotherapy.

Machine Learning-Based Identification of Diagnostic Biomarkers for Korean Male Sarcopenia Through Integrative DNA Methylation and Methylation Risk Score: From the Korean Genomic Epidemiology Study (KoGES).

Sarcopenia, characterized by a progressive decline in muscle mass, strength, and function, is primarily attributable to aging. DNA methylation, influenced by both genetic predispositions and environmental exposures, plays a significant role in sarcopenia occurrence. This study employed machine learning (ML) methods to identify differentially methylated probes (DMPs) capable of diagnosing sarcopenia in middle-aged individuals. We also investigated the relationship between muscle strength, muscle mass, age, and sarcopenia risk as reflected in methylation profiles. Data from 509 male participants in the urban cohort of the Korean Genome Epidemiology Study_Health Examinee study were categorized into quartile groups based on the sarcopenia criteria for appendicular skeletal muscle index (ASMI) and handgrip strength (HG). To identify diagnostic biomarkers for sarcopenia, we used recursive feature elimination with cross validation (RFECV), to pinpoint DMPs significantly associated with sarcopenia. An ensemble model, leveraging majority voting, was utilized for evaluation. Furthermore, a methylation risk score (MRS) was calculated, and its correlation with muscle strength, function, and age was assessed using likelihood ratio analysis and multinomial logistic regression. Participants were classified into two groups based on quartile thresholds: sarcopenia (n = 37) with ASMI and HG in the lowest quartile, and normal ranges (n = 48) in the highest. In total, 238 DMPs were identified and eight probes were selected using RFECV. These DMPs were used to build an ensemble model with robust diagnostic capabilities for sarcopenia, as evidenced by an area under the receiver operating characteristic curve of 0.94. Based on eight probes, the MRS was calculated and then validated by analyzing age, HG, and ASMI among the control group (n = 424). Age was positively correlated with high MRS (coefficient, 1.2494; odds ratio [OR], 3.4882), whereas ASMI and HG were negatively correlated with high MRS (ASMI coefficient, -0.4275; OR, 0.6521; HG coefficient, -0.3116; OR, 0.7323). Overall, this study identified key epigenetic markers of sarcopenia in Korean males and developed a ML model with high diagnostic accuracy for sarcopenia. The MRS also revealed significant correlations between these markers and age, HG, and ASMI. These findings suggest that both diagnostic models and the MRS can play an important role in managing sarcopenia in middle-aged populations.

RAIN: machine learning-based identification for HIV-1 bNAbs

Broadly neutralizing antibodies (bNAbs) are promising candidates for the treatment and prevention of HIV-1 infections. Despite their critical importance, automatic detection of HIV-1 bNAbs from immune repertoires is still lacking. Here, we develop a straightforward computational method for the Rapid Automatic Identification of bNAbs (RAIN) based on machine learning methods. In contrast to other approaches, which use one-hot encoding amino acid sequences or structural alignment for prediction, RAIN uses a combination of selected sequence-based features for the accurate prediction of HIV-1 bNAbs. We demonstrate the performance of our approach on non-biased, experimentally obtained and sequenced BCR repertoires from HIV-1 immune donors. RAIN processing leads to the successful identification of distinct HIV-1 bNAbs targeting the CD4-binding site of the envelope glycoprotein. In addition, we validate the identified bNAbs using an in vitro neutralization assay and we solve the structure of one of them in complex with the soluble native-like heterotrimeric envelope glycoprotein by single-particle cryo-electron microscopy (cryo-EM). Overall, we propose a method to facilitate and accelerate HIV-1 bNAbs discovery from non-selected immune repertoires.

An Enhanced Modeling Framework for Bearing Fault Simulation and Machine Learning-based Identification with Bayesian-optimized Hyperparameter Tuning

Abstract Condition monitoring of rotating machinery offers a salient tool for predictive maintenance on rolling elements, subjected to continuous working loads, wear, fatigue, and degradation. In this study, an enhanced computational tool for bearing fault simulation and feature extraction is proposed. A subsequent identification scheme is realized, through Bayesian optimization of hyperparameters, including support vector classifier (SVC), gradient boosting (GBoost), random forest (RF), extreme gradient boosting (XBoost), light gradient boosting machine (LightGBM), and categorical boosting (CatBoost). The proposed hyperparameter optimization technique stands out from traditional methods by offering a more informed and efficient pathway to optimal performance in predictive maintenance. Utilizing Bayesian optimization for hyperparameter tuning of machine learning models, which has not been extensively explored in this field, our approach demonstrates significant advancements. Typical instances of bearing faults, like inner race, outer race, and ball faults are considered. The analysis relies on extraction of statistical and engineering features form collected response signals, including kurtosis, root mean square, peak, and ridge factor. Highly influential variables are highlighted based on feature selection and importance algorithms, allowing bearing fault classification. We demonstrate that SVC and LightGBM produce over 97% of accuracy at low computational cost. This approach constitutes a scalable and robust framework for similar applications in engineering diagnostics.

Machine learning-based identification and immune characterization of ferroptosis-related molecular clusters in osteoarthritis and validation.

Osteoarthritis (OA), a degenerative joint disease, involves synovial inflammation, subchondral bone erosion, and cartilage degeneration. Ferroptosis, a regulated non-apoptotic programmed cell death, is associated with various diseases. This study investigates ferroptosis-related molecular subtypes in OA to comprehend underlying mechanisms. The Gene Expression Omnibus datasets GSE206848, GSE55457, GSE55235, GSE77298 and GSE82107 were used utilized. Unsupervised clustering identified the ferroptosis-related gene (FRG) subtypes, and their immune characteristics were assessed. FRG signatures were derived using LASSO and SVM-RFE algorithms, forming models to evaluate OA's ferroptosis-related immune features. Three FRG clusters were found to be immunologically heterogeneous, with cluster 1 displaying robust immune response. Models identified nine key signature genes via algorithms, demonstrating strong diagnostic and prognostic performance. Finally, qRT-PCR and Western blot validated these genes, offering consistent results. In addition, some of these genes may have implications as new therapeutic targets and can be used to guide clinical applications.

Machine Learning-Based Identification of Mating Type and Metalaxyl Response in Phytophthora infestans Using SSR Markers.

Phytophthora infestans is the causal agent of late blight in potato. The occurrence of P. infestans with both A1 and A2 mating types in the field may result in sexual reproduction and the generation of recombinant strains. Such strains with new combinations of traits can be highly aggressive, resistant to fungicides, and can make the disease difficult to control in the field. Metalaxyl-resistant isolates are now more prevalent in potato fields. Understanding the genetic structure and rapid identification of mating types and metalaxyl response of P. infestans in the field is a prerequisite for effective late blight disease monitoring and management. Molecular and phenotypic assays involving molecular and phenotypic markers such as mating types and metalaxyl response are typically conducted separately in the studies of the genotypic and phenotypic diversity of P. infestans. As a result, there is a pressing need to reduce the experimental workload and more efficiently assess the aggressiveness of different strains. We think that employing genetic markers to not only estimate genotypic diversity but also to identify the mating type and fungicide response using machine learning techniques can guide and speed up the decision-making process in late blight disease management, especially when the mating type and metalaxyl resistance data are not available. This technique can also be applied to determine these phenotypic traits for dead isolates. In this study, over 600 P. infestans isolates from different populations-Estonia, Pskov region, and Poland-were classified for mating types and metalaxyl response using machine learning techniques based on simple sequence repeat (SSR) markers. For both traits, random forest and the support vector machine demonstrated good accuracy of over 70%, compared to the decision tree and artificial neural network models whose accuracy was lower. There were also associations (p < 0.05) between the traits and some of the alleles detected, but machine learning prediction techniques based on multilocus SSR genotypes offered better prediction accuracy.