An explainable machine learning-driven proposal of pulmonary fibrosis biomarkers.
Ontology highlight
ABSTRACT:
SUBMITTER: Fanidis D
PROVIDER: S-EPMC10049879 | biostudies-literature | 2023
REPOSITORIES: biostudies-literature
Publications
An explainable machine learning-driven proposal of pulmonary fibrosis biomarkers.
Fanidis Dionysios D Pezoulas Vasileios C VC Fotiadis Dimitrios Ι DΙ Aidinis Vassilis V
Computational and structural biotechnology journal 20230325
Pulmonary fibrosing diseases are in the very epicenter of biomedical research both due to their increasing prevalence and their association with SARS-CoV-2 infections. Research of idiopathic pulmonary fibrosis, the most lethal among the interstitial lung diseases, is in need for new biomarkers and potential disease targets, a goal that could be accelerated using machine learning techniques. In this study, we have used Shapley values to explain the decisions made by an ensemble learning model tra ...[more]
Similar Datasets
Project description:Idiopathic pulmonary fibrosis (IPF) is the most common and serious type of idiopathic interstitial pneumonia, characterized by chronic, progressive, and low survival rates, while unknown disease etiology. Until recently, patients with idiopathic pulmonary fibrosis have a poor prognosis, high mortality, and limited treatment options, due to the lack of effective early diagnostic and prognostic tools. Therefore, we aimed to identify biomarkers for idiopathic pulmonary fibrosis based on multiple machine-learning approaches and to evaluate the role of immune infiltration in the disease. The gene expression profile and its corresponding clinical data of idiopathic pulmonary fibrosis patients were downloaded from Gene Expression Omnibus (GEO) database. Next, the differentially expressed genes (DEGs) with the threshold of FDR < 0.05 and |log2 foldchange (FC)| > 0.585 were analyzed via R package "DESeq2" and GO enrichment and KEGG pathways were run in R software. Then, least absolute shrinkage and selection operator (LASSO) logistic regression, support vector machine-recursive feature elimination (SVM-RFE) and random forest (RF) algorithms were combined to screen the key potential biomarkers of idiopathic pulmonary fibrosis. The diagnostic performance of these biomarkers was evaluated through receiver operating characteristic (ROC) curves. Moreover, the CIBERSORT algorithm was employed to assess the infiltration of immune cells and the relationship between the infiltrating immune cells and the biomarkers. Finally, we sought to understand the potential pathogenic role of the biomarker (SLAIN1) in idiopathic pulmonary fibrosis using a mouse model and cellular model. A total of 3658 differentially expressed genes of idiopathic pulmonary fibrosis were identified, including 2359 upregulated genes and 1299 downregulated genes. FHL2, HPCAL1, RNF182, and SLAIN1 were identified as biomarkers of idiopathic pulmonary fibrosis using LASSO logistic regression, RF, and SVM-RFE algorithms. The ROC curves confirmed the predictive accuracy of these biomarkers both in the training set and test set. Immune cell infiltration analysis suggested that patients with idiopathic pulmonary fibrosis had a higher level of B cells memory, Plasma cells, T cells CD8, T cells follicular helper, T cells regulatory (Tregs), Macrophages M0, and Mast cells resting compared with the control group. Correlation analysis demonstrated that FHL2 was significantly associated with the infiltrating immune cells. qPCR and western blotting analysis suggested that SLAIN1 might be a signature for the diagnosis of idiopathic pulmonary fibrosis. In this study, we identified four potential biomarkers (FHL2, HPCAL1, RNF182, and SLAIN1) and evaluated the potential pathogenic role of SLAIN1 in idiopathic pulmonary fibrosis. These findings may have great significance in guiding the understanding of disease mechanisms and potential therapeutic targets in idiopathic pulmonary fibrosis.
| S-EPMC10545744 | biostudies-literature
Project description:IntroductionIdiopathic pulmonary fibrosis (IPF) is a severe chronic respiratory disease characterized by treatment challenges and poor prognosis. Identifying relevant biomarkers for effective early-stage risk prediction is therefore of critical importance.MethodsIn this study, we obtained gene expression profiles and corresponding clinical data of IPF patients from the GEO database. GO enrichment and KEGG pathway analyses were performed using R software. To construct an IPF risk prediction model, we employed LASSO-Cox regression analysis and the SVM-RFE algorithm. PODNL1 and PIGA were identified as potential biomarkers associated with IPF onset, and their predictive accuracy was confirmed using ROC curve analysis in the test set. Furthermore, GSEA revealed enrichment in multiple pathways, while immune function analysis demonstrated a significant correlation between IPF onset and immune cell infiltration. Finally, the roles of PODNL1 and PIGA as biomarkers were validated through in vivo and in vitro experiments using qRT-PCR, Western blotting, and immunohistochemistry.ResultsThese findings suggest that PODNL1 and PIGA may serve as critical biomarkers for IPF onset and contribute to its pathogenesis.DiscussionThis study highlights their potential for early biomarker discovery and risk prediction in IPF, offering insights into disease mechanisms and diagnostic strategies.
| S-EPMC11811625 | biostudies-literature
Project description:ObjectiveThis study aimed to identify candidate gene biomarkers associated with immune infiltration in idiopathic pulmonary fibrosis (IPF) based on machine learning algorithms.MethodsMicroarray datasets of IPF were extracted from the Gene Expression Omnibus (GEO) database to screen for differentially expressed genes (DEGs). The DEGs were subjected to enrichment analysis, and two machine learning algorithms were used to identify candidate genes associated with IPF. These genes were verified in a validation cohort from the GEO database. Receiver operating characteristic (ROC) curves were plotted to assess the predictive value of the IPF-associated genes. The cell-type identification by estimating relative subsets of RNA transcripts (CIBERSORT) algorithm was used to evaluate the proportion of immune cells in IPF and normal tissues. Additionally, the correlation between the expression of IPF-associated genes and the infiltration levels of immune cells was examined.ResultsA total of 302 upregulated and 192 downregulated genes were identified. Functional annotation, pathway enrichment, Disease Ontology and gene set enrichment analyses revealed that the DEGs were related to the extracellular matrix and immune responses. COL3A1, CDH3, CEBPD, and GPIHBP1 were identified as candidate biomarkers using machine learning algorithms, and their predictive value was verified in a validation cohort. Additionally, ROC analysis revealed that the four genes had high predictive accuracy. The infiltration levels of plasma cells, M0 macrophages and resting dendritic cells were higher and those of resting natural killer (NK) cells, M1 macrophages and eosinophils were lower in the lung tissues of patients with IPF than in those of healthy individuals. The expression of the abovementioned genes was correlated with the infiltration levels of plasma cells, M0 macrophages and eosinophils.ConclusionCOL3A1, CDH3, CEBPD, and GPIHBP1 are candidate biomarkers of IPF. Plasma cells, M0 macrophages and eosinophils may be involved in the development of IPF and may serve as immunotherapeutic targets in IPF.
| S-EPMC9968813 | biostudies-literature
Project description:BackgroundIdiopathic Pulmonary Fibrosis (IPF) presents a severe respiratory challenge with a poor prognosis due to the lack of reliable biomarkers. Recent evidence suggests that Endoplasmic Reticulum Stress (ERS) may be associated with IPF pathogenesis. This study focuses on uncovering ERS-associated biomarkers for IPF.MethodsSequencing data from diverse datasets were analyzed, utilizing differential gene expression analysis and Weighted Gene Co-expression Network Analysis (WGCNA). Endoplasmic Reticulum Stress (ERS)-related genes were extracted from the GeneCards database. Hub genes were identified through Protein-Protein Interaction (PPI) analysis. Diagnostic and prognostic models were developed using machine learning algorithms and validated across both training and validation sets. Additionally, techniques such as Cell-type Identification by Estimating Relative Subsets of RNA Transcripts and single-cell RNA sequencing were employed to identify potential IPF-related cells. These findings were further investigated to elucidate their underlying mechanisms through in vitro experiments.ResultsDifferentially expressed genes, WGCNA-identified blue module genes, and ERS-related genes extracted from the GeneCards database were intersected, and the resulting genes were used to construct diagnostic and prognostic models. Validation using multiple datasets indicated that both the diagnostic and prognostic models possess strong predictive capabilities. PPI analysis highlighted SPP1 as a potential hub gene in IPF. Moreover, M2 macrophages were found in higher quantities in the lung tissue of IPF patients, with a significant increase in SPP1-expressing M2 macrophages compared to the control group. In vitro experiments demonstrated that exogenous SPP1 inhibited the proliferation and migration of M2 macrophages and promoted apoptosis within a certain concentration range.ConclusionThis study identifies ERS-related biomarkers in IPF, highlighting SPP1 and M2 macrophages. The resulting diagnostic and prognostic models offer strong predictive capabilities, unveiling new therapeutic avenues.
| S-EPMC11441044 | biostudies-literature
Project description:BackgroundDiabetic retinopathy (DR), a prevalent complication in patients with type 2 diabetes, has attracted increasing attention. Recent studies have explored a plausible association between retinopathy and significant liver fibrosis. The aim of this investigation was to develop a sophisticated machine learning (ML) model, leveraging comprehensive clinical datasets, to forecast the likelihood of significant liver fibrosis in patients with retinopathy and to interpret the ML model by applying the SHapley Additive exPlanations (SHAP) method.MethodsThis inquiry was based on data from the National Health and Nutrition Examination Survey 2005-2008 cohort. Utilizing the Fibrosis-4 index (FIB-4), liver fibrosis was stratified across a spectrum of grades (F0-F4). The severity of retinopathy was determined using retinal imaging and segmented into four discrete gradations. A ten-fold cross-validation approach was used to gauge the propensity towards liver fibrosis. Eight ML methodologies were used: Extreme Gradient Boosting, Random Forest, multilayer perceptron, Support Vector Machines, Logistic Regression (LR), Plain Bayes, Decision Tree, and k-nearest neighbors. The efficacy of these models was gauged using metrics, such as the area under the curve (AUC). The SHAP method was deployed to unravel the intricacies of feature importance and explicate the inner workings of the ML model.ResultsThe analysis included 5,364 participants, of whom 2,116 (39.45%) exhibited notable liver fibrosis. Following random allocation, 3,754 individuals were assigned to the training set and 1,610 were allocated to the validation cohort. Nine variables were curated for integration into the ML model. Among the eight ML models scrutinized, the LR model attained zenith in both AUC (0.867, 95% CI: 0.855-0.878) and F1 score (0.749, 95% CI: 0.732-0.767). In internal validation, this model sustained its superiority, with an AUC of 0.850 and an F1 score of 0.736, surpassing all other ML models. The SHAP methodology unveils the foremost factors through importance ranking.ConclusionSophisticated ML models were crafted using clinical data to discern the propensity for significant liver fibrosis in patients with retinopathy and to intervene early.Practice implicationsImproved early detection of liver fibrosis risk in retinopathy patients enhances clinical intervention outcomes.
| S-EPMC11552118 | biostudies-literature
Project description:Selecting the right individuals for a sports team, organization, or military unit has a large influence on the achievements of the organization. However, the approaches commonly used for selection are either not reporting predictive performance or not explainable (i.e., black box models). In the present study, we introduce a novel approach to selection research, using various machine learning models. We examined 274 special forces recruits, of whom 196 dropped out, who performed a set of physical and psychological tests. On this data, we compared four machine learning models on their predictive performance, explainability, and stability. We found that a stable rule-based (SIRUS) model was most suitable for classifying dropouts from the special forces selection program. With an averaged area under the curve score of 0.70, this model had good predictive performance, while remaining explainable and stable. Furthermore, we found that both physical and psychological variables were related to dropout. More specifically, a higher score on the 2800 m time, need for connectedness, and skin folds was most strongly associated with dropping out. We discuss how researchers and practitioners can benefit from these insights in sport and performance contexts.
| S-EPMC11534633 | biostudies-literature
Project description:BackgroundShort-term readmission for pediatric pulmonary hypertension (PH) is associated with a substantial social and personal burden. However, tools to predict individualized readmission risk are lacking. This study aimed to develop machine learning models to predict 30-day unplanned readmission in children with PH.MethodsThis study collected data on pediatric inpatients with PH from the Chongqing Medical University Medical Data Platform from January 2012 to January 2019. Key clinical variables were selected by the least absolute shrinkage and the selection operator. Prediction models were selected from 15 machine learning algorithms with excellent performance, which was evaluated by area under the operating characteristic curve (AUC). The outcome of the predictive model was interpreted by SHapley Additive exPlanations (SHAP).ResultsA total of 5,913 pediatric patients with PH were included in the final cohort. The CatBoost model was selected as the predictive model with the greatest AUC for 0.81 (95% CI: 0.77-0.86), high accuracy for 0.74 (95% CI: 0.72-0.76), sensitivity 0.78 (95% CI: 0.69-0.87), and specificity 0.74 (95% CI: 0.72-0.76). Age, length of stay (LOS), congenital heart surgery, and nonmedical order discharge showed the greatest impact on 30-day readmission in pediatric PH, according to SHAP results.ConclusionsThis study developed a CatBoost model to predict the risk of unplanned 30-day readmission in pediatric patients with PH, which showed more significant performance compared with traditional logistic regression. We found that age, LOS, congenital heart surgery, and nonmedical order discharge were important factors for 30-day readmission in pediatric PH.
| S-EPMC9360407 | biostudies-literature
Project description:The study developed accurate explainable machine learning (ML) models for predicting first-time acute exacerbation of chronic obstructive pulmonary disease (COPD, AECOPD) at an individual level. We conducted a retrospective case-control study. A total of 606 patients with COPD were screened for eligibility using registry data from the COPD Pay-for-Performance Program (COPD P4P program) database at Changhua Christian Hospital between January 2017 and December 2019. Recursive feature elimination technology was used to select the optimal subset of features for predicting the occurrence of AECOPD. We developed four ML models to predict first-time AECOPD, and the highest-performing model was applied. Finally, an explainable approach based on ML and the SHapley Additive exPlanations (SHAP) and a local explanation method were used to evaluate the risk of AECOPD and to generate individual explanations of the model's decisions. The gradient boosting machine (GBM) and support vector machine (SVM) models exhibited superior discrimination ability (area under curve [AUC] = 0.833 [95% confidence interval (CI) 0.745-0.921] and AUC = 0.836 [95% CI 0.757-0.915], respectively). The decision curve analysis indicated that the GBM model exhibited a higher net benefit in distinguishing patients at high risk for AECOPD when the threshold probability was <0.55. The COPD Assessment Test (CAT) and the symptom of wheezing were the two most important features and exhibited the highest SHAP values, followed by monocyte count and white blood cell (WBC) count, coughing, red blood cell (RBC) count, breathing rate, oral long-acting bronchodilator use, chronic pulmonary disease (CPD), systolic blood pressure (SBP), and others. Higher CAT score; monocyte, WBC, and RBC counts; BMI; diastolic blood pressure (DBP); neutrophil-to-lymphocyte ratio; and eosinophil and lymphocyte counts were associated with AECOPD. The presence of symptoms (wheezing, dyspnea, coughing), chronic disease (CPD, congestive heart failure [CHF], sleep disorders, and pneumonia), and use of COPD medications (triple-therapy long-acting bronchodilators, short-acting bronchodilators, oral long-acting bronchodilators, and antibiotics) were also positively associated with AECOPD. A high breathing rate, heart rate, or systolic blood pressure and methylxanthine use were negatively correlated with AECOPD. The ML model was able to accurately assess the risk of AECOPD. The ML model combined with SHAP and the local explanation method were able to provide interpretable and visual explanations of individualized risk predictions, which may assist clinical physicians in understanding the effects of key features in the model and the model's decision-making process.
| S-EPMC8879653 | biostudies-literature
Project description:Patients with advanced idiopathic pulmonary fibrosis (IPF), a complex and incurable lung disease with an elusive pathology, are nearly exclusive candidates for lung transplantation. Improved identification of patient subtypes can enhance early diagnosis and intervention, ultimately leading to better prognostic outcomes for patients. The goal of this study is to identify new immune patterns and biomarkers in patients. Immune subtypes in IPF patients were identified using single-sample gene set enrichment analysis, and immune subtype-related genes were explored using the weighted correlation network analysis algorithm. A machine learning integration framework was used to establish the optimal prognostic model, known as the immune-related risk score (IRS). Single-cell sequencing was conducted to investigate the major role of macrophage-derived PLA2G7 in the immune microenvironment. We assessed the stability of celecoxib in targeting PLA2G7 through molecular docking and surface plasmon resonance. IPF patients present two distinct immune subtypes, one characterized by immune activation and inflammation, and the other by immune suppression. IRS can predict the immune status and prognosis of IPF patients. Furthermore, multi-cohort analysis and single-cell sequencing analysis demonstrated the diagnostic and prognostic value of PLA2G7 derived from macrophages and its role in shaping the inflammatory immune microenvironment in IPF patients. Celecoxib could effectively and stably bind with PLA2G7. PLA2G7, as identified through IRS, demonstrates marked stability in diagnosing and predicting the prognosis of IPF patients as well as predicting their immune status. It can serve as a novel biomarker for IPF patients.
| S-EPMC11437001 | biostudies-literature
Project description:Change in chromosome copy number, or aneuploidy, is a typical feature of cancer cells, which can alter the protein abundance of hundreds or thousands of proteins. Yet, evidence from aneuploid cell lines suggests that the protein dosage changes are to a great extent buffered, and thus the abundance of many proteins is similar to the abundance in euploid cells despite gene copy number changes. Despite ubiquitous occurrence of protein dosage buffering, there is limited understanding of the involved molecular mechanisms. Moreover, whether protein dosage buffering affects all proteins and cells to similar degree, and whether it brings adaptive advantage remains unclear. Here, we established a novel approach to quantify protein dosage buffering in a gene copy-number dependent manner, and show that dosage compensation is wide-spread in cancer cell lines and in-vivo tumor samples, yet the extent of buffering is variable. By devising multifactorial machine learning (ML) models that efficiently predict gene dosage buffering and the usage of explainable ML techniques, we demonstrate that mean gene dependency, protein complex participation, haploinsufficiency and mRNA decay are strongly predictive for buffering. Our analysis also indicates that proteotoxic stress and drug sensitivity is reduced in samples that exhibit high protein buffering on average.
2025-09-12 | PXD060017 | Pride