Random forest algorithm-based classification model of pesticide aquatic toxicity to fishes

Machine Learning-Driven Identification of Exosome-Related Genes in Head and Neck Squamous Cell Carcinoma for Prognostic Evaluation and Drug Response Prediction

Background: This study integrated four Gene Expression Omnibus (GEO) datasets to identify disease-specific feature genes in head and neck squamous cell carcinoma (HNSCC) through differential expression analysis with batch effect correction. Methods: The GeneCards database was used to find genes related to exosomes, and samples were categorized into groups with high and low expression levels based on these feature genes. Functional and pathway enrichment analyses (GO, KEGG, and GSEA) were used to investigate the possible biological mechanisms underlying feature genes. A predictive model was produced by using machine learning algorithms (LASSO regression, SVM, and random forest) to find disease-specific feature genes. Receiver operating characteristic (ROC) curve analysis was used to assess the model's effectiveness. The diagnostic model showed excellent predictive accuracy through external data GSE83519 validation. Results: This analysis highlighted 22 genes with significant differential expression. A predictive model based on five important genes (AGRN, TSPAN6, MMP9, HBA1, and PFN2) was produced by using machine learning algorithms. MMP9 and TSPAN6 showed relatively high predictive performance. Using the ssGSEA algorithm, three key genes (MMP9, AGRN, and PFN2) were identified as strongly linked to immune regulation, immune response suppression, and critical signaling pathways involved in HNSCC progression. Matching HNSCC feature gene expression profiles with DSigDB compound signatures uncovered potential therapeutic targets. Molecular docking simulations identified ligands with high binding affinity and stability, notably C5 and Hoechst 33258, which were prioritized for further validation and potential drug development. Conclusions: This study employs a novel diagnostic model for HNSCC constructed using machine learning technology, which can provide support for the early diagnosis of HNSCC and thus contribute to improving patient treatment plans and clinical management strategies.

Predicting acute toxicity of pesticides towards Daphnia magna with random forest algorithm

A large number of pesticides are released into the environment, resulting in serious threat for aquatic organisms. In this work, 15 quantum chemical descriptors were used to develop a quantitative structure-activity/toxicity relationship (QSAR/QSTR) model for toxicity pEC50 of 745 pesticides towards Daphnia magna, by using random forest algorithm. The optimal QSTR model in this paper yielded a coefficient of determination of 0.828, root-mean-square error of 0.798, and mean absolute error of 0.628 for the test set of 149 pesticides, which are accurate values compared with those of QSTR models published recently. Research has revealed that increasing molecular size (or molar volume), the most positive atomic Mulliken (or APT) charge with hydrogens summed into heavy, and the highest occupied molecular orbital (HOMO) energy, can result in higher toxicity pEC50. Increasing the lowest unoccupied molecular orbital (LUMO) energy and the HOMO and LUMO energy gap can lead to lower toxicity pEC50.

Pesticides in fish from the Uruguay River and markets in Argentina and health risk assessment

Organochlorine, organophosphate, triazole, and strobilurin pesticides were determined in fish samples. Relative standard deviations lower than 9.3% were obtained for organochlorine pesticides and 10.8% for other pesticides. Accuracy ranged from 73% to 119% for organochlorine pesticides and 80.4% to 116% for organophosphate, triazole, and strobilurin pesticides. A total of 28 pesticides were analysed and 7 of them were detected (exceeding 10 µg/kg) in some samples, with the highest concentration recorded at 68.5 µg/kg, corresponding to heptachlor epoxide A. The pesticide most frequently detected was β HCH, found in 30 of the 100 analysed samples. Hazard Quotient values were estimated for men, women, and children. These values exceeded 1 for heptachlor epoxide in women and children, as well as for endrin in children. These findings emphasise the need for stricter controls to reduce fish contamination and mitigate health risks.

Screening the Best Risk Model and Susceptibility SNPs for Chronic Obstructive Pulmonary Disease (COPD) Based on Machine Learning Algorithms

Background and Purpose

Chronic obstructive pulmonary disease (COPD) is a common and progressive disease that is influenced by both genetic and environmental factors, and genetic factors are important determinants of COPD. This study focuses on screening the best predictive models for assessing COPD-associated SNPs and then using the best models to predict potential risk factors for COPD.

Methods

Healthy subjects (n=290) and COPD patients (n=233) were included in this study, the Agena MassARRAY platform was applied to genotype the subjects for SNPs. The selected sample loci were first screened by logistic regression analysis, based on which the key SNPs were further screened by LASSO regression, RFE algorithm and Random Forest algorithm, and the ROC curves were plotted to assess the discriminative performance of the models to screen the best prediction model. Finally, the best prediction model was used for the prediction of risk factors for COPD.

Results

One-way logistic regression analysis screened 44 candidate SNPs from 146 SNPs, on the basis of which 44 SNPs were screened or feature ranked using LASSO model, RFE-Caret, RFE-Lda, RFE-lr, RFE-nb, RFE-rf, RFE-treebag algorithms and random forest model, respectively, and obtained ROC curve values of 0.809, 0.769, 0.798, 0.743, 0.686, 0.766, 0.743, 0.719, respectively, so we selected the lasso model as the best model, and then constructed a column-line graph model for the 25 SNPs screened in it, and found that rs12479210 might be the potential risk factors for COPD.

Conclusion

The LASSO model is the best predictive model for COPD and rs12479210 may be a potential risk locus for COPD.

Quantitative structure-activity relationship predicting toxicity of pesticides towards Daphnia magna

Global pesticide usage reaching 2.7 million metric tons annually, brings a grave threat to non-target organisms, especially aquatic organisms, resulting in serious concerns. Predicting aquatic toxicity of pesticides towards Daphnia magna is significant. In this work, random forest (RF) algorithm, together with ten Dragon molecular descriptors, was successfully utilized to develop a quantitative structure-activity/toxicity relationship (QSAR/QSTR) model for the toxicity pEC50 of 745 pesticides towards Daphnia magna. The optimal QSTR model (RF Model I) based on the RF parameters of ntree = 50, mtry = 3 and nodesize = 5, yielded R2 = 0.877, MAE = 0.570, rms = 0.739 (training set of 596 pEC50), R2 = 0.807, MAE = 0.732, rms = 0.902 (test set of 149 pEC50), and R2 = 0.863, MAE = 0.602, rms = 0.774 (total set of 745 pEC50), which are accurate and satisfactory. The optimal RF model is comparable to other published QSTR models for Daphnia magna, although the optimal RF model possessed a small descriptor subset and dealt with a large dataset of pesticide toxicity pEC50. Thus, the investigation in this work provides a reliable, applicable QSTR model for predicting the toxicity pEC50 of pesticides towards Daphnia magna.

Classification Model of Pesticide Toxicity in Americamysis bahia Based on Quantum Chemical Descriptors

A set of quantum chemical descriptors (molecular polarization, heat capacity, entropy, Mulliken net charge of the most positive hydrogen atom, APT charge of the most negative atom and APT charge of the most positive atom with hydrogen summed into heavy atoms) was successfully used to establish the classification models for the toxicity pLC50 of pesticides in Americamysis bahia. The optimal random forest model (Class Model A) yielded predictive accuracy of 100% (training set of 217 pesticides), 95.8% (test set of 72 pesticides) and 99.0% (total set of 289 pesticides), which were very satisfactory, compared with previous classification models reported for the toxicity of compounds in aquatic organisms. Therefore, it is reasonable to apply the quantum chemical descriptors associated with molecular structural information on molecular bulk, chemical reactivity and weak interactions, to develop classification models for the toxicity pLC50 of pesticides in A. bahia.

Finding key genes (UBE2T, KIF4A, CDCA3, and CDCA5) co-expressed in hepatitis, cirrhosis and hepatocellular carcinoma based on multiple bioinformatics techniques

BACKGROUND

Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide. Hepatitis B virus (HBV) is one of the major causes of liver cirrhosis (LC) and HCC. Therefore, the discovery of common markers for hepatitis B or LC and HCC is crucial for the prevention of HCC.

METHODS

Expressed genes for to chronic active hepaititis B (CAH-B), LC and HCC were obtained from the GEO and TCGA databases, and co-expressed genes were screened using Protein-protein interaction (PPI) networks, least absolute shrinkage and selection operator (LASSO), random forest (RF) and support vector machine - recursive feature elimination (SVM-RFE). The prognostic value of genes was assessed using Kaplan-Meier (KM) survival curves. Columnar line plots, calibration curves and receiver operating characteristic (ROC) curves of individual genes were used for evaluation. Validation was performed using GEO datasets. The association of these key genes with HCC clinical features was explored using the UALCAN database ( https://ualcan.path.uab.edu/index.html ).

RESULTS

Based on WGCNA analysis and TCGA database, the co-expressed genes (565) were screened. Moreover, the five algorithms of MCODE (ClusteringCoefficient, MCC, Degree, MNC, and DMNC) was used to select one of the most important and most closely linked clusters (the top 50 genes ranked). Using, LASSO regression model, RF model and SVM-RFE model, four key genes (UBE2T, KIF4A, CDCA3, and CDCA5) were identified for subsequent research analysis. These 4 genes were highly expressed and associated with poor prognosis and clinical features in HCC patients.

CONCLUSION

These four key genes (UBE2T, KIF4A, CDCA3, and CDCA5) may be common biomarkers for CAH-B and HCC or LC and HCC, promising to advance our understanding of the molecular basis of CAH-B/LC/HCC progression.

Quantitative structure-activity relationship model development for estimating the predicted No-effect concentration of petroleum hydrocarbon and derivatives in the ecological risk assessment

Quantitative structure-activity relationship (QSAR) is a cost-effective solution to directly and accurately estimating the environmental safety thresholds (ESTs) of pollutants in the ecological risk assessment due to the lack of toxicity data. In this study, QSAR models were developed for estimating the Predicted No-Effect Concentrations (PNECs) of petroleum hydrocarbons and their derivatives (PHDs) under dietary exposure, based on the quantified molecular descriptors and the obtained PNECs of 51 PHDs with given acute or chronic toxicity concentrations. Three high-reliable QSAR models were respectively developed for PHDs, aromatic hydrocarbons and their derivatives (AHDs), and alkanes, alkenes and their derivatives (ALKDs), with excellent fitting performance evidenced by high correlation coefficient (0.89-0.95) and low root mean square error (0.13-0.2 mg/kg), and high stability and predictive performance reflected by high internal and external verification coefficient (Q2LOO, 0.66-0.89; Q2F1, 0.62-0.78; Q2F2, 0.60-0.73). The investigated quantitative relationships between molecular structure and PNECs indicated that 18 autocorrelation descriptors, 3 information index descriptors, 4 barysz matrix descriptors, 6 burden modified eigenvalues descriptors, and 1 BCUT descriptor were important molecular descriptors affecting the PNECs of PHDs. The obtained results supported that PNECs of PHDs can be accurately estimated by the influencing molecular descriptors and the quantitative relationship from the developed QSAR models, that provided a new feasible solution for ESTs derivation in the ecological risk assessment.

Gut microbiota landscape and potential biomarker identification in female patients with systemic lupus erythematosus using machine learning

Objectives

Systemic Lupus Erythematosus (SLE) is a complex autoimmune disease that disproportionately affects women. Early diagnosis and prevention are crucial for women's health, and the gut microbiota has been found to be strongly associated with SLE. This study aimed to identify potential biomarkers for SLE by characterizing the gut microbiota landscape using feature selection and exploring the use of machine learning (ML) algorithms with significantly dysregulated microbiotas (SDMs) for early identification of SLE patients. Additionally, we used the SHapley Additive exPlanations (SHAP) interpretability framework to visualize the impact of SDMs on the risk of developing SLE in females.

Methods

Stool samples were collected from 54 SLE patients and 55 Negative Controls (NC) for microbiota analysis using 16S rRNA sequencing. Feature selection was performed using Elastic Net and Boruta on species-level taxonomy. Subsequently, four ML algorithms, namely logistic regression (LR), Adaptive Boosting (AdaBoost), Random Forest (RF), and eXtreme gradient boosting (XGBoost), were used to achieve early identification of SLE with SDMs. Finally, the best-performing algorithm was combined with SHAP to explore how SDMs affect the risk of developing SLE in females.

Results

Both alpha and beta diversity were found to be different in SLE group. Following feature selection, 68 and 21 microbiota were retained in Elastic Net and Boruta, respectively, with 16 microbiota overlapping between the two, i.e., SDMs for SLE. The four ML algorithms with SDMs could effectively identify SLE patients, with XGBoost performing the best, achieving Accuracy, Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value, and AUC values of 0.844, 0.750, 0.938, 0.923, 0.790, and 0.930, respectively. The SHAP interpretability framework showed a complex non-linear relationship between the relative abundance of SDMs and the risk of SLE, with Escherichia_fergusonii having the largest SHAP value.

Conclusions

This study revealed dysbiosis in the gut microbiota of female SLE patients. ML classifiers combined with SDMs can facilitate early identification of female patients with SLE, particularly XGBoost. The SHAP interpretability framework provides insight into the impact of SDMs on the risk of SLE and may inform future scientific treatment for SLE.

Global classification models for predicting acute toxicity of chemicals towards Daphnia magna

Molecular descriptors reflecting structural information on hydrophobicity, reactivity, polarizability, hydrogen bond and charged groups, were used to predict the toxicity (pLC50) of chemicals towards Daphnia magna with global quantitative structure-activity/toxicity relationship (QSAR/QSTR) models. A sufficiently large dataset including 1517 chemical toxicity to Daphnia magna was divided into a training set (758 pLC50) and a test set (759 pLC50). By applying random forest algorithm, two classification models, Class Model A and Class Model B were developed, having prediction accuracy, sensitivity and specificity above 85% for Class 1 (with pLC50 ≤ 4.48) and Class 2 (with pLC50 > 4.48). The Class Model A was based on nine molecular descriptors and RF parameters of nodesize = 1, ntree = 80 and mtry = 2, and yielded accuracy of 92.3% (training set), 85.6% (test set) and 88.9% (total data set). Class Model B was based on ten descriptors and parameters, nodesize = 1, ntree = 90 and mtry = 2, produced accuracy of 88.3% (training set), 86.8% (test set) and 87.5% (total data set). The two classification models were satisfactory compared with other classification model reported in the literature, although classification models in this work dealt with more samples. Thus, the two classification models with a larger applicability domain provided efficient tools for assessing chemical aquatic toxicity towards Daphnia magna.

Quantitative Read-across structure-activity relationship (q-RASAR): A new approach methodology to model aquatic toxicity of organic pesticides against different fish species

We have developed quantitative toxicity prediction models for organic pesticides of agricultural importance considering different fish species using a novel quantitative Read-across structure-activity relationship (q-RASAR) approach. The current study uses experimental (Log 1/LC50) data of organic pesticides to various fish species, including Rainbow trout (RT: Oncorhynchus mykiss: 715 data points), Lepomis (LP: Lepomis macrochirus: 136 data points), and Miscellaneous (Pimephales promelas, Brachydanio rerio: 226 data points). This study has also discussed the validation of the developed models and the analysis of structural features that are important for aquatic toxicity towards fishes. The read-across-derived similarity, error, and concordance measures (RASAR descriptors) have been extracted from the preliminary 0D-2D descriptors; the combined pool of RASAR and selected 0D-2D descriptors have been used to develop the final models by employing partial least squares algorithm. All the q-RASAR models are acceptable in terms of goodness of fit, robustness, and external predictivity, superseding the quality of the respective QSAR models, as seen from the computed validation metrics. The q-RASAR is an effective approach that has the potential to be used as a good alternative way to enhance external predictivity, interpretability, and transferability for aquatic toxicity prediction as well as ecotoxicity potential identification.

Mitigation of avermectin exposure-induced brain tissue damage in carp by quercetin

Avermectin is widely used as an important insecticide in agricultural production, but it also shows strong toxicity to non-target organisms. Quercetin is a natural flavonoid that is widely used due to its good anti-inflammatory and antioxidant effects. We believe that quercetin may have a potential therapeutic effect on avermectin poisoning. This experiment was proposed to observe the effect of quercetin on the toxic response to avermectin by observing the toxic response caused by avermectin in the brain of carp. In this project, 60 carp were studied as control group (Control), quercetin administration group (QUE), avermectin exposure group (AVM) and quercetin treatment avermectin exposure group (QUE + AVM) with different interventions to study the effect of quercetin on avermectin. The carp brain tissues were stained and simultaneously analyzed for blood-brain barrier (BBB), oxidative stress indicators, inflammatory factors, and apoptosis using qPCR technique. The results of the study indicate that avermectin exhibits a neurotoxic mechanism of action in fish by decreasing the transcript levels of tight junction protein-related genes, which in turn leads to the rupture of the BBB in the carp brain tissue. Avermectin induced apoptosis in carp brain tissue by increasing oxidative stress response and promoting inflammatory cell infiltration. Quercetin could reduce the accumulation of reactive oxygen species (ROS) in the brain tissue of carp caused by avermectin exposure toxicity, maintain redox homeostasis, reduce inflammatory response, and protect brain tissue cells from apoptosis. The present study confirmed the therapeutic and protective effects of quercetin on neurotoxicity in carp caused by avermectin exposure.

What is the ecotoxicity of a given chemical for a given aquatic species? Predicting interactions between species and chemicals using recommender system techniques

Ecotoxicological safety assessment of chemicals requires toxicity data on multiple species, despite the general desire of minimizing animal testing. Predictive models, specifically machine learning (ML) methods, are one of the tools capable of solving this apparent contradiction as they allow to generalize toxicity patterns across chemicals and species. However, despite the availability of large public toxicity datasets, the data is highly sparse, complicating model development. The aim of this study is to provide insights into how ML can predict toxicity using a large but sparse dataset. We developed models to predict LC50-values, based on experimental LC50-data covering 2431 organic chemicals and 1506 aquatic species from the ECOTOX-database. Several well-known ML techniques were evaluated and a new ML model was developed, inspired by recommender systems. This new model involves a simple linear model that learns low-rank interactions between species and chemicals using factorization machines. We evaluated the predictive performances of the developed models based on two validation settings: 1) predicting unseen chemical-species pairs, and 2) predicting unseen chemicals. The results of this study show that ML models can accurately predict LC50-values in both validation settings. Moreover, we show that the novel factorization machine approach can match well-tuned, complex, ML approaches.

Hybrid system model for wind abrasion segmentation using semi-automatic classification of remote sensing multispectral areas

Wind abrasion, caused by particles transported by strong winds impacting on structures, can lead to their degradation. Although this phenomenon has hardly been studied in this context, it is becoming increasingly important due to new trends in infrastructure location, especially in renewable energy terms. Metallic structures are particularly vulnerable to degradation by the action of windblown sand particles. However, characterising such secluded sites is complicated, and remote sensing systems and satellite information become crucial. The objective of this research is to identify and delineate the geographic areas that are vulnerable to this phenomenon by employing a hybrid model with historical data and the semi-automatic classification of multispectral satellite images. The model is based on critical variables identified by the scientific community and case studies documented in the literature. The methodology used for the study consists of four phases, including creating a scientifically robust database, downloading and managing satellite and historical long-term information, segmenting the regions of interest, and modelling using supervised classification techniques. The proposed algorithm shows very accurate results (R2 = 0.9922) and the overall system approach is presented as a useful and generalizable method to address this problem, increasing the existing knowledge on material wear by particle action, and contributing to optimizing the initial design of resilient structures.

Identification of immunogenic cell death-related signature on prognosis and immunotherapy in kidney renal clear cell carcinoma

Background

Immunogenic cell death (ICD) is considered a particular cell death modality of regulated cell death (RCD) and plays a significant role in various cancers. The connection between kidney renal clear cell carcinoma (KIRC) and ICD remains to be thoroughly explored.

Methods

We conducted a variety of bioinformatics analyses using R software, including cluster analysis, prognostic analysis, enrichment analysis and immune infiltration analysis. In addition, we performed Quantitative Real-time PCR to evaluate RNA levels of specific ICD genes. The proliferation was measured through Cell Counting Kit-8 (CCK-8) assay and colony-formation assay in RCC cell lines.

Results

We determined two ICD subtypes through consensus clustering analysis. The two subtypes showed significantly different clinical outcomes, genomic alterations and tumor immune microenvironment. Moreover, we constructed the ICD prognostic signature based on TF, FOXP3, LY96, SLC7A11, HSP90AA1, UCN, IFNB1 and TLR3 and calculated the risk score for each patient. Kaplan-Meier survival analysis and ROC curve demonstrated that patients in the high-risk group had significantly poorer prognosis compared with the low-risk group. We then validated the signature through external cohort and further evaluated the relation between the signature and clinical features, tumor immune microenvironment and immunotherapy response. Given its critical role in ICD, we conducted further analysis on LY96. Our results indicated that downregulation of LY96 inhibited the proliferation ability of RCC cells.

Conclusions

Our research revealed the underlying function of ICD in KIRC and screened out a potential biomarker, which provided a novel insight into individualized immunotherapy in KIRC.

Large Dataset-Based Regression Model of Chemical Toxicity to Vibrio fischeri

For the first time, a global regression quantitative structure-toxicity/activity relationship (QSTR/QSAR) model was developed for the toxicity of a large data set including 1236 chemicals towards Vibrio fischeri, by using random forest (RF) regression algorithm. The optimal RF model with RF parameters of mtry = 3, ntree = 150 and nodesize = 5 was based on 13 molecular descriptors. It can achieve accurate prediction for the toxicity of 99.1% of 1236 chemicals, and yield coefficients of determination R2 of 0.893 for 930 log(Mw/IBC50) in the training set, 0.723 for 306 log(Mw/IBC50) in the test se, and 0.865 for 1236 toxicity log(Mw/IBC50) in the total set. The optimal RF global model proposed in this work is comparable to other published local QSTR models on small datasets of the toxicity to Vibrio fischeri.

Support Vector Machine-Based Global Classification Model of the Toxicity of Organic Compounds to Vibrio fischeri

Vibrio fischeri is widely used as the model species in toxicity and risk assessment. For the first time, a global classification model was proposed in this paper for a two-class problem (Class − 1 with log1/IBC50 ≤ 4.2 and Class + 1 with log1/IBC50 > 4.2, the unit of IBC50: mol/L) by utilizing a large data set of 601 toxicity log1/IBC50 of organic compounds to Vibrio fischeri. Dragon software was used to calculate 4885 molecular descriptors for each compound. Stepwise multiple linear regression (MLR) analysis was used to select the descriptor subset for the models. The ten molecular descriptors used in the classification model reflect the structural information on the Michael-type addition of nucleophiles, molecular branching, molecular size, polarizability, hydrophobic, and so on. Furthermore, these descriptors were interpreted from the point of view of toxicity mechanisms. The optimal support vector machine (SVM) model (C = 253.8 and γ = 0.009) was obtained with the genetic algorithm. The SVM classification model produced a prediction accuracy of 89.1% for the training set (451 log1/IBC50), of 80.0% for the test set (150 log1/IBC50), and of 86.9% for the total data set (601 log1/IBC50), which are higher than that (80.5%, 76%, and 79.4%, respectively) from the binary logistic regression (BLR) model. The global SVM classification model is successful, although it deals with a large data set in relation to the toxicity of organics to Vibrio fischeri.

Quantitative relationships between national cultures and the increase in cases of novel coronavirus pneumonia

Support vector machine (SVM) and genetic algorithm were successfully used to predict the changes in the prevalence rate (ΔPR) measured by the increase of reported cases per million population from the 16th to the 45th day during a nation's lockdown after the COVID-19 outbreak. The national cultural indices [individualism-collectivism (Ind), tightness-looseness (Tight)], and the number of people per square kilometer (Pop_density) were used to develop the SVM model of lnΔPR. The SVM model has R² of 0.804 for the training set (44 samples) and 0.853 for the test set (11 samples), which were much higher than those (0.416 and 0.593) of the multiple linear regression model. The statistical results indicate that there are nonlinear relationships between lnΔPR and Tight, Ind, and Pop_density. It is feasible to build the model for lnΔPR with SVM algorithm. The results suggested that the risk of COVID-19 epidemic spread will be reduced if a nation implements severe measures to strengthen the tightness of national culture and individuals realize the importance of collectivism.