Identification of endoplasmic reticulum stress-associated genes and subtypes for prediction of Alzheimer’s disease based on interpretable machine learning

Interpreting artificial intelligence models: a systematic review on the application of LIME and SHAP in Alzheimer's disease detection

Explainable artificial intelligence (XAI) has gained much interest in recent years for its ability to explain the complex decision-making process of machine learning (ML) and deep learning (DL) models. The Local Interpretable Model-agnostic Explanations (LIME) and Shaply Additive exPlanation (SHAP) frameworks have grown as popular interpretive tools for ML and DL models. This article provides a systematic review of the application of LIME and SHAP in interpreting the detection of Alzheimer's disease (AD). Adhering to PRISMA and Kitchenham's guidelines, we identified 23 relevant articles and investigated these frameworks' prospective capabilities, benefits, and challenges in depth. The results emphasise XAI's crucial role in strengthening the trustworthiness of AI-based AD predictions. This review aims to provide fundamental capabilities of LIME and SHAP XAI frameworks in enhancing fidelity within clinical decision support systems for AD prognosis.

Mag-Net: Rapid enrichment of membrane-bound particles enables high coverage quantitative analysis of the plasma proteome

Membrane-bound particles in plasma are composed of exosomes, microvesicles, and apoptotic bodies and represent ~1-2% of the total protein composition. Proteomic interrogation of this subset of plasma proteins augments the representation of tissue-specific proteins, representing a "liquid biopsy," while enabling the detection of proteins that would otherwise be beyond the dynamic range of liquid chromatography-tandem mass spectrometry of unfractionated plasma. We have developed an enrichment strategy (Mag-Net) using hyper-porous strong-anion exchange magnetic microparticles to sieve membrane-bound particles from plasma. The Mag-Net method is robust, reproducible, inexpensive, and requires <100 μL plasma input. Coupled to a quantitative data-independent mass spectrometry analytical strategy, we demonstrate that we can collect results for >37,000 peptides from >4,000 plasma proteins with high precision. Using this analytical pipeline on a small cohort of patients with neurodegenerative disease and healthy age-matched controls, we discovered 204 proteins that differentiate (q-value < 0.05) patients with Alzheimer's disease dementia (ADD) from those without ADD. Our method also discovered 310 proteins that were different between Parkinson's disease and those with either ADD or healthy cognitively normal individuals. Using machine learning we were able to distinguish between ADD and not ADD with a mean ROC AUC = 0.98 ± 0.06.

An interactive web application to identify early Parkinsonian non-tremor-dominant subtypes

BACKGROUND

Parkinson's disease (PD) patients with tremor-dominant (TD) and non-tremor-dominant (NTD) subtypes exhibit heterogeneity. Rapid identification of different motor subtypes may help to develop personalized treatment plans.

METHODS

The data were acquired from the Parkinson's Disease Progression Marker Initiative (PPMI). Following the identification of predictors utilizing recursive feature elimination (RFE), seven classical machine learning (ML) models, including logistic regression, support vector machine, decision tree, random forest, extreme gradient boosting, etc., were trained to predict patients' motor subtypes, evaluating the performance of models through the area under the receiver operating characteristic curve (AUC) and validating by the follow-up data.

RESULTS

The feature subset engendered by RFE encompassed 20 features, comprising some clinical assessments and cerebrospinal fluid α-synuclein (CSF α-syn). ML models fitted in the RFE subset performed better in the test and validation sets. The best performing model was support vector machines with the polynomial kernel (P-SVM), achieving an AUC of 0.898. Five-fold repeated cross-validation showed the P-SVM model with CSF α-syn performed better than the model without CSF α-syn (P = 0.034). The Shapley additive explanation plot (SHAP) illustrated that how the levels of each feature affect the predicted probability as NTD subtypes.

CONCLUSION

An interactive web application was developed based on the P-SVM model constructed from feature subset by RFE. It can identify the current motor subtypes of PD patients, making it easier to understand the status of patients and develop personalized treatment plans.

Comprehensive analysis of an endoplasmic reticulum stress-related gene prediction model and immune infiltration in idiopathic pulmonary fibrosis

Background

Idiopathic pulmonary fibrosis (IPF) is a chronic progressive interstitial lung disease. This study aimed to investigate the involvement of endoplasmic reticulum stress (ERS) in IPF and explore its correlation with immune infiltration.

Methods

ERS-related differentially expressed genes (ERSRDEGs) were identified by intersecting differentially expressed genes (DEGs) from three Gene Expression Omnibus datasets with ERS-related gene sets. Gene Set Variation Analysis and Gene Ontology were used to explore the potential biological mechanisms underlying ERS. A nomogram was developed using the risk signature derived from the ERSRDEGs to perform risk assessment. The diagnostic value of the risk signature was evaluated using receiver operating characteristics, calibration, and decision curve analyses. The ERS score of patients with IPF was measured using a single-sample Gene Set Enrichment Analysis (ssGSEA) algorithm. Subsequently, a prognostic model based on the ERS scores was established. The proportion of immune cell infiltration was assessed using the ssGSEA and CIBERSORT algorithms. Finally, the expression of ERSRDEGs was validated in vivo and in vitro via RT-qPCR.

Results

This study developed an 8-ERSRDEGs signature. Based on the expression of these genes, we constructed a diagnostic nomogram model in which agouti-related neuropeptide had a significantly greater impact on the model. The area under the curve values for the predictive value of the ERSRDEGs signature were 0.975 and 1.000 for GSE70866 and GSE110147, respectively. We developed a prognostic model based on the ERS scores of patients with IPF. Furthermore, we classified patients with IPF into two subtypes based on their signatures. The RT-qPCR validation results supported the reliability of most of our conclusions.

Conclusion

We developed and verified a risk model using eight ERSRDEGs. These eight genes can potentially affect the progression of IPF by regulating ERS and immune responses.

Identifying Explainable Machine Learning Models and a Novel SFRP2+ Fibroblast Signature as Predictors for Precision Medicine in Ovarian Cancer

Ovarian cancer (OC) is a type of malignant tumor with a consistently high mortality rate. The diagnosis of early-stage OC and identification of functional subsets in the tumor microenvironment are essential to the development of patient management strategies. However, the development of robust models remains unsatisfactory. We aimed to utilize artificial intelligence and single-cell analysis to address this issue. Two independent datasets were screened from the Gene Expression Omnibus (GEO) database and processed to obtain overlapping differentially expressed genes (DEGs) in stage II-IV vs. stage I diseases. Three explainable machine learning algorithms were integrated to construct models that could determine the tumor stage and extract important characteristic genes as diagnostic biomarkers. Correlations between cancer-associated fibroblast (CAF) infiltration and characteristic gene expression were analyzed using TIMER2.0 and their relationship with survival rates was comprehensively explored via the Kaplan-Meier plotter (KM-plotter) online database. The specific expression of characteristic genes in fibroblast subsets was investigated through single-cell analysis. A novel fibroblast subset signature was explored to predict immune checkpoint inhibitor (ICI) response and oncogene mutation through Tumor Immune Dysfunction and Exclusion (TIDE) and artificial neural network algorithms, respectively. We found that Support Vector Machine-Shapley Additive Explanations (SVM-SHAP), Extreme Gradient Boosting (XGBoost), and Random Forest (RF) successfully diagnosed early-stage OC (stage I). The area under the receiver operating characteristic curves (AUCs) of these models exceeded 0.990. Their overlapping characteristic gene, secreted frizzled-related protein 2 (SFRP2), was a risk factor that affected the overall survival of OC patients with stage II-IV disease (log-rank test: p < 0.01) and was specifically expressed in a fibroblast subset. Finally, the SFRP2+ fibroblast signature served as a novel predictor in evaluating ICI response and exploring pan-cancer tumor protein P53 (TP53) mutation (AUC = 0.853, 95% confidence interval [CI]: 0.829-0.877). In conclusion, the models based on SVM-SHAP, XGBoost, and RF enabled the early detection of OC for clinical decision making, and SFRP2+ fibroblast signature used in diagnostic models can inform OC treatment selection and offer pan-cancer TP53 mutation detection.

A glance through the effects of CD4+ T cells, CD8+ T cells, and cytokines on Alzheimer's disease

Alzheimer's disease (AD) is the most common form of dementia. Unfortunately, despite numerous studies, an effective treatment for AD has not yet been established. There is remarkable evidence indicating that the innate immune mechanism and adaptive immune response play significant roles in the pathogenesis of AD. Several studies have reported changes in CD8+ and CD4+ T cells in AD patients. This mini-review article discusses the potential contribution of CD4+ and CD8+ T cells reactivity to amyloid β (Aβ) protein in individuals with AD. Moreover, this mini-review examines the potential associations between T cells, heme oxygenase (HO), and impaired mitochondria in the context of AD. While current mathematical models of AD have not extensively addressed the inclusion of CD4+ and CD8+ T cells, there exist models that can be extended to consider AD as an autoimmune disease involving these T cell types. Additionally, the mini-review covers recent research that has investigated the utilization of machine learning models, considering the impact of CD4+ and CD8+ T cells.

Machine learning identification and immune infiltration of disulfidptosis-related Alzheimer's disease molecular subtypes

BACKGROUND

Alzheimer's disease (AD) is a common neurodegenerative disorder. Disulfidptosis is a newly discovered form of programmed cell death that holds promise as a therapeutic strategy for various disorders. However, the functional roles of disulfidptosis-related genes (DRGs) in AD remain unknown.

METHODS

Microarray data and clinical information from patients with AD and healthy controls were downloaded from the Gene Expression Omnibus database. A thorough examination of DRG expression and immune characteristics in both groups was performed. Based on the identified DRGs, we performed an unsupervised clustering analysis to categorize the AD samples into various disulfidptosis-related molecular clusters. Weighted gene co-expression network analysis was performed to select hub genes specific to disulfidptosis-related AD clusters. The performances of various machine learning models were compared to determine the optimal predictive model. The predictive ability of the optimal model was assessed using nomogram analysis and five external datasets.

RESULTS

Eight DRGs showed differential expression between the AD and control samples. Two different molecular clusters were identified. The immune cell infiltration analysis revealed distinct differences in the immune microenvironment of the two clusters. The support vector machine model showed the highest performance, and a panel of five signature genes was identified, which showed excellent performance on the external validation datasets. The nomogram analysis also showed high accuracy in predicting AD.

CONCLUSION

We identified disulfidptosis-related molecular clusters in AD and established a novel risk model to assess the likelihood of developing AD. These findings revealed a complex association between disulfidptosis and AD, which may aid in identifying potential therapeutic targets for this debilitating disorder.

Identification of metabolism-related subtypes and feature genes in Alzheimer's disease

BACKGROUND

Owing to the heterogeneity of Alzheimer's disease (AD), its pathogenic mechanisms are yet to be fully elucidated. Evidence suggests an important role of metabolism in the pathophysiology of AD. Herein, we identified the metabolism-related AD subtypes and feature genes.

METHODS

The AD datasets were obtained from the Gene Expression Omnibus database and the metabolism-relevant genes were downloaded from a previously published compilation. Consensus clustering was performed to identify the AD subclasses. The clinical characteristics, correlations with metabolic signatures, and immune infiltration of the AD subclasses were evaluated. Feature genes were screened using weighted correlation network analysis (WGCNA) and processed via Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analyses. Furthermore, three machine-learning algorithms were used to narrow down the selection of the feature genes. Finally, we identified the diagnostic value and expression of the feature genes using the AD dataset and quantitative reverse-transcription polymerase chain reaction (qRT-PCR) analysis.

RESULTS

Three AD subclasses were identified, namely Metabolism Correlated (MC) A (MCA), MCB, and MCC subclasses. MCA contained signatures associated with high AD progression and may represent a high-risk subclass compared with the other two subclasses. MCA exhibited a high expression of genes related to glycolysis, fructose, and galactose metabolism, whereas genes associated with the citrate cycle and pyruvate metabolism were downregulated and associated with high immune infiltration. Conversely, MCB was associated with citrate cycle genes and exhibited elevated expression of immune checkpoint genes. Using WGCNA, 101 metabolic genes were identified to exhibit the strongest association with poor AD progression. Finally, the application of machine-learning algorithms enabled us to successfully identify eight feature genes, which were employed to develop a nomogram model that could bring distinct clinical benefits for patients with AD. As indicated by the AD datasets and qRT-PCR analysis, these genes were intimately associated with AD progression.

CONCLUSION

Metabolic dysfunction is associated with AD. Hypothetical molecular subclasses of AD based on metabolic genes may provide new insights for developing individualized therapy for AD. The feature genes highly correlated with AD progression included GFAP, CYB5R3, DARS, KIAA0513, EZR, KCNC1, COLEC12, and TST.

Construction of a neural network diagnostic model for renal fibrosis and investigation of immune infiltration characteristics

Background

Recently, the incidence rate of renal fibrosis has been increasing worldwide, greatly increasing the burden on society. However, the diagnostic and therapeutic tools available for the disease are insufficient, necessitating the screening of potential biomarkers to predict renal fibrosis.

Methods

Using the Gene Expression Omnibus (GEO) database, we obtained two gene array datasets (GSE76882 and GSE22459) from patients with renal fibrosis and healthy individuals. We identified differentially expressed genes (DEGs) between renal fibrosis and normal tissues and analyzed possible diagnostic biomarkers using machine learning. The diagnostic effect of the candidate markers was evaluated using receiver operating characteristic (ROC) curves and verified their expression using Reverse transcription quantitative polymerase chain reaction (RT-qPCR). The CIBERSORT algorithm was used to determine the proportions of 22 types of immune cells in patients with renal fibrosis, and the correlation between biomarker expression and the proportion of immune cells was studied. Finally, we developed an artificial neural network model of renal fibrosis.

Results

Four candidate genes namely DOCK2, SLC1A3, SOX9 and TARP were identified as biomarkers of renal fibrosis, with the area under the ROC curve (AUC) values higher than 0.75. Next, we verified the expression of these genes by RT-qPCR. Subsequently, we revealed the potential disorder of immune cells in the renal fibrosis group through CIBERSORT analysis and found that immune cells were highly correlated with the expression of candidate markers.

Conclusion

DOCK2, SLC1A3, SOX9, and TARP were identified as potential diagnostic genes for renal fibrosis, and the most relevant immune cells were identified. Our findings provide potential biomarkers for the diagnosis of renal fibrosis.

Identification of immune microenvironment subtypes and signature genes for Alzheimer's disease diagnosis and risk prediction based on explainable machine learning

Background

Using interpretable machine learning, we sought to define the immune microenvironment subtypes and distinctive genes in AD.

Methods

ssGSEA, LASSO regression, and WGCNA algorithms were used to evaluate immune state in AD patients. To predict the fate of AD and identify distinctive genes, six machine learning algorithms were developed. The output of machine learning models was interpreted using the SHAP and LIME algorithms. For external validation, four separate GEO databases were used. We estimated the subgroups of the immunological microenvironment using unsupervised clustering. Further research was done on the variations in immunological microenvironment, enhanced functions and pathways, and therapeutic medicines between these subtypes. Finally, the expression of characteristic genes was verified using the AlzData and pan-cancer databases and RT-PCR analysis.

Results

It was determined that AD is connected to changes in the immunological microenvironment. WGCNA revealed 31 potential immune genes, of which the greenyellow and blue modules were shown to be most associated with infiltrated immune cells. In the testing set, the XGBoost algorithm had the best performance with an AUC of 0.86 and a P-R value of 0.83. Following the screening of the testing set by machine learning algorithms and the verification of independent datasets, five genes (CXCR4, PPP3R1, HSP90AB1, CXCL10, and S100A12) that were closely associated with AD pathological biomarkers and allowed for the accurate prediction of AD progression were found to be immune microenvironment-related genes. The feature gene-based nomogram may provide clinical advantages to patients. Two immune microenvironment subgroups for AD patients were identified, subtype2 was linked to a metabolic phenotype, subtype1 belonged to the immune-active kind. MK-866 and arachidonyltrifluoromethane were identified as the top treatment agents for subtypes 1 and 2, respectively. These five distinguishing genes were found to be intimately linked to the development of the disease, according to the Alzdata database, pan-cancer research, and RT-PCR analysis.

Conclusion

The hub genes associated with the immune microenvironment that are most strongly associated with the progression of pathology in AD are CXCR4, PPP3R1, HSP90AB1, CXCL10, and S100A12. The hypothesized molecular subgroups might offer novel perceptions for individualized AD treatment.