Mechanisms Of Complex Diseases Research Articles

Clinical Assessment of Osteoarthritis Pain: Contemporary Scenario, Challenges, and Future Perspectives.

The multifaceted nature of osteoarthritis (OA) pain presents a challenge in understanding and managing the condition. The diverse pain experiences, progression rates, individual responses to treatments, and complex disease mechanisms contribute to heterogeneity in the clinical studies outcomes. The lack of a standardized methodology for assessing and classifying OA pain challenges healthcare practitioners. This complicates the establishment of universally applicable protocols or standardized guidelines for treatment. This article explores the heterogeneity observed in clinical studies evaluating OA pain treatments, highlighting the necessity for refined methodologies, personalized patient categorization, and consistent outcome measures. It discusses the role of the multidimensional nature of OA pain, underlying pain mechanisms, and other contributing factors to the heterogeneity in outcome measures. Addressing these variations is crucial to establishing a more consistent framework for evidence-based treatments and advancing care of the patient with OA pain.

Chronic Pancreatitis-Associated Metabolic Bone Diseases: Epidemiology, Mechanisms, and Clinical Advances.

Chronic pancreatitis (CP) is a progressive inflammatory disease with an increasing global prevalence. In recent years, a strong association between CP and metabolic bone diseases (MBDs), especially osteoporosis, has been identified, attracting significant attentions in the research field. Epidemiological data suggest a rising trend in the incidence of MBDs among CP patients. Notably, recent studies have highlighted a profound interplay between CP and altered nutritional and immune profiles, offering insights into its linkage with MBDs.At molecular level, CP introduces a series of biochemical disturbances that compromise bone homeostasis. One critical observation is the disrupted metabolism of vitamin D and vitamin K, both essential micronutrients for maintaining bone integrity, in CP patients. In this review, we provide physio-pathological perspectives on the development and mechanisms of CP-related MBDs. We also outline some of the latest therapeutic strategies for treating patients with CP-associated MBDs, including stem cell transplantation, monoclonal antibodies, and probiotic therapy. In summary, CP-associated MBDs represent a rising medical challenge, involving multiple tissues and organs, complex disease mechanisms, and diverse treatment approaches. More in-depth studies are required for understanding the complex interplay between CP and MBDs to facilitate the development of more specific and effective therapeutic approaches.

Pathway Activation Analysis for Pan-Cancer Personalized Characterization Based on Riemannian Manifold.

The pathogenesis of carcinoma is believed to come from the combined effect of polygenic variation, and the initiation and progression of malignant tumors are closely related to the dysregulation of biological pathways. Quantifying the alteration in pathway activation and identifying coordinated patterns of pathway dysfunction are the imperative part of understanding the malignancy process and distinguishing different tumor stages or clinical outcomes of individual patients. In this study, we have conducted in silico pathway activation analysis using Riemannian manifold (RiePath) toward pan-cancer personalized characterization, which is the first attempt to apply the Riemannian manifold theory to measure the extent of pathway dysregulation in individual patient on the tangent space of the Riemannian manifold. RiePath effectively integrates pathway and gene expression information, not only generating a relatively low-dimensional and biologically relevant representation, but also identifying a robust panel of biologically meaningful pathway signatures as biomarkers. The pan-cancer analysis across 16 cancer types reveals the capability of RiePath to evaluate pathway activation accurately and identify clinical outcome-related pathways. We believe that RiePath has the potential to provide new prospects in understanding the molecular mechanisms of complex diseases and may find broader applications in predicting biomarkers for other intricate diseases.

Modelling the micro- and macro- environment of pancreatic cancer: from patients to pre-clinical models and back.

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy with very low survival rates. Over the past 50 years, improvements in PDAC survival have significantly lagged behind the progress made in other cancers. PDAC's dismal prognosis is due to typical late-stage diagnosis combined with lack of effective treatments and complex mechanisms of disease. We propose that improvements in survival are partly hindered by the current focus on largely modelling and targeting PDAC as one disease, despite it being heterogeneous. Implementing new disease-representative pre-clinical mouse models that capture this complexity could enable the development of transformative therapies. Specifically, these models should recapitulate human PDAC late-stage biology, heterogeneous genetics, extensive non-malignant stroma, and associated risk factors and comorbidities. In this Perspective, we focus on how pre-clinical mouse models could be improved to exemplify key features of PDAC micro- and macro- environments, which would drive clinically relevant patient stratification, tailored treatments and improved survival.

Abstract 6237: Large-scale imputation models for multi-ancestry proteome-wide association analysis

Abstract Proteome-wide association studies (PWAS) decode the intricate proteomic landscape of biological mechanisms for complex diseases. Traditional PWAS model training relies heavily on individual-level reference proteomes, thereby restricting its capacity to harness the emerging summary-level protein quantitative trait loci (pQTL) data in the public domain. Here we introduced a novel framework to train PWAS models directly from pQTL summary statistics. By leveraging extensive pQTL data from the UK Biobank, deCODE, and ARIC studies, we applied our approach to train large-scale European PWAS models (total n = 88,838 subjects). Furthermore, we developed PWAS models tailored for Asian and African ancestries by integrating multi-ancestry summary and individual-level data resources (total n = 914 for Asian and 3,042 for African ancestries). We validated the performance of our PWAS models through a systematic multi-ancestry analysis of over 700 phenotypes across five major genetic data resources. Specifically, we identified several novel and well-known genes for various cancers; we highlighted eight genes associated with six cancer traits across all three PWAS model cohorts, including CTSF, RSPO3, CRTAM, LAYN, CHRDL2, DPEP1, NFASC, and NAAA. Our results bridge the gap between genomics and proteomics for drug discovery, highlighting novel protein-phenotype links and their transferability across diverse ancestries. The developed PWAS models and data resources are freely available at www.gcbhub.org. Citation Format: Chong Wu, Zichen Zhang, Xiaochen Yang, Bingxin Zhao. Large-scale imputation models for multi-ancestry proteome-wide association analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6237.

A critical appreciation of pathway analysis in atherosclerotic disease. Cellular phenotypic plasticity as an illustrative example

The rapid advancements in genome-scale (omics) techniques has created significant opportunities to investigate complex disease mechanisms in tissues and cells. Nevertheless, interpreting -omics data can be challenging, and pathway enrichment analysis is a frequently used method to identify candidate molecular pathways that drive gene expression changes. With a growing number of -omics studies dedicated to atherosclerosis, there has been a significant increase in studies and hypotheses relying on enrichment analysis.This brief review discusses the benefits and limitations of pathway enrichment analysis within atherosclerosis research. We highlight the challenges of identifying complex biological processes, such as cell phenotypic switching, within -omics data. Additionally, we emphasize the need for more comprehensive and curated gene sets that reflect the biological complexity of atherosclerosis.Pathway enrichment analysis is a valuable tool for gaining insights into the molecular mechanisms of atherosclerosis. Nevertheless, it is crucial to remain aware of the intrinsic limitations of this approach. By addressing these weaknesses, enrichment analysis in atherosclerosis can lead to breakthroughs in identifying the mechanisms of disease progresses, the identification of key driver genes, and consequently, advance personalized patient care.

A comprehensive database of exosome molecular biomarkers and disease-gene associations

Exosomes play a crucial role in intercellular communication and can be used as biomarkers for diagnostic and therapeutic clinical applications. However, systematic studies in cancer-associated exosomal nucleic acids remain a big challenge. Here, we developed ExMdb, a comprehensive database of exosomal nucleic acid biomarkers and disease-gene associations curated from published literature and high-throughput datasets. We performed a comprehensive curation of exosome properties including 4,586 experimentally supported gene-disease associations, 13,768 diagnostic and therapeutic biomarkers, and 312,049 nucleic acid subcellular locations. To characterize expression variation of exosomal molecules and identify causal factors of complex diseases, we have also collected 164 high-throughput datasets, including bulk and single-cell RNA sequencing (scRNA-seq) data. Based on these datasets, we performed various bioinformatics and statistical analyses to support our conclusions and advance our knowledge of exosome biology. Collectively, our dataset will serve as an essential resource for investigating the regulatory mechanisms of complex diseases and improving the development of diagnostic and therapeutic biomarkers.

Integrating joint latent class mixed models and Bayesian network for uncovering clinical subgroups of COVID-19 patients

When modelling the dynamics of biomarkers in biomedical studies, it is essential to identify homogeneous clusters of patients and analyse them from a precision medicine perspective. This need has emerged as crucial and urgent during the COVID-19 pandemic: early understanding of symptoms and patient heterogeneity has significant implications for prevention, early diagnosis, effective management, and treatment. Additionally, biomarker progression may be associated with clinically relevant time-toevent data. Therefore, statistical models are necessary to gain insight into complex disease mechanisms by properly accounting for unobservable heterogeneity in patients while jointly modelling longitudinal and time-to-event data. In this study, we leverage the key features of Latent Class modelling and Bayesian Network approaches and propose a unified framework to (a) uncover homogeneous subgroups of patients concerning their longitudinal and survival data and (b) describe patient subgroups within a multivariate framework.

Integrated causal inference modeling uncovers novel causal factors and potential therapeutic targets of Qingjin Yiqi granules for chronic fatigue syndrome

Objective: Chronic fatigue syndrome (CFS) is a prevalent symptom of post-coronavirus disease 2019 (COVID-19) and is associated with unclear disease mechanisms. The herbal medicine Qingjin Yiqi granules (QJYQ) constitute a clinically approved formula for treating post-COVID-19; however, its potential as a drug target for treating CFS remains largely unknown. This study aimed to identify novel causal factors for CFS and elucidate the potential targets and pharmacological mechanisms of action of QJYQ in treating CFS. Methods: This prospective cohort analysis included 4,212 adults aged ≥65 years who were followed up for 7 years with 435 incident CFS cases. Causal modeling and multivariate logistic regression analysis were performed to identify the potential causal determinants of CFS. A proteome-wide, two-sample Mendelian randomization (MR) analysis was employed to explore the proteins associated with the identified causal factors of CFS, which may serve as potential drug targets. Furthermore, we performed a virtual screening analysis to assess the binding affinity between the bioactive compounds in QJYQ and CFS-associated proteins. Results: Among 4,212 participants (47.5% men) with a median age of 69 years (interquartile range: 69–70 years) enrolled in 2004, 435 developed CFS by 2011. Causal graph analysis with multivariate logistic regression identified frequent cough (odds ratio: 1.74, 95% confidence interval [CI]: 1.15–2.63) and insomnia (odds ratio: 2.59, 95% CI: 1.77–3.79) as novel causal factors of CFS. Proteome-wide MR analysis revealed that the upregulation of endothelial cell-selective adhesion molecule (ESAM) was causally linked to both chronic cough (odds ratio: 1.019, 95% CI: 1.012–1.026, P = 2.75 e−05) and insomnia (odds ratio: 1.015, 95% CI: 1.008–1.022, P = 4.40 e−08) in CFS. The major bioactive compounds of QJYQ, ginsenoside Rb2 (docking score: −6.03) and RG4 (docking score: −6.15), bound to ESAM with high affinity based on virtual screening. Conclusions: Our integrated analytical framework combining epidemiological, genetic, and in silico data provides a novel strategy for elucidating complex disease mechanisms, such as CFS, and informing models of action of traditional Chinese medicines, such as QJYQ. Further validation in animal models is warranted to confirm the potential pharmacological effects of QJYQ on ESAM and as a treatment for CFS.

DDK-Linker: a network-based strategy identifies disease signals by linking high-throughput omics datasets to disease knowledge.

The high-throughput genomic and proteomic scanning approaches allow investigators to measure the quantification of genome-wide genes (or gene products) for certain disease conditions, which plays an essential role in promoting the discovery of disease mechanisms. The high-throughput approaches often generate a large gene list of interest (GOIs), such as differentially expressed genes/proteins. However, researchers have to perform manual triage and validation to explore the most promising, biologically plausible linkages between the known disease genes and GOIs (disease signals) for further study. Here, to address this challenge, we proposed a network-based strategy DDK-Linker to facilitate the exploration of disease signals hidden in omics data by linking GOIs to disease knowns genes. Specifically, it reconstructed gene distances in the protein-protein interaction (PPI) network through six network methods (random walk with restart, Deepwalk, Node2Vec, LINE, HOPE, Laplacian) to discover disease signals in omics data that have shorter distances to disease genes. Furthermore, benefiting from the establishment of knowledge base we established, the abundant bioinformatics annotations were provided for each candidate disease signal. To assist in omics data interpretation and facilitate the usage, we have developed this strategy into an application that users can access through a website or download the R package. We believe DDK-Linker will accelerate the exploring of disease genes and drug targets in a variety of omics data, such as genomics, transcriptomics and proteomics data, and provide clues for complex disease mechanism and pharmacological research. DDK-Linker is freely accessible at http://ddklinker.ncpsb.org.cn/.

Abstract C043: Spatial proteomic immune profiling of pancreatic ductal adenocarcinoma and the tumor microenvironment

Abstract Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal disease with poor prognosis and treatment options. Technological challenges in dissecting the immune constituents of the PDAC tumor microenvironment (TME) has limited our understanding of the complex disease mechanisms in vivo, which may in part explain the lack of novel therapeutics for this recalcitrant disease. Here, we performed spatial proteomics using the multiplexed ion beam imaging (MIBI) approach to quantify 36 spatial signatures in tissues from a cohort of twelve treatment naïve PDAC patients. We applied a deep learning segmentation algorithm to better resolve expression levels at the single cell resolution, and robustly stratify phenotypes through an iterative process of clustering and annotation, with expert visual validation. The combination of phenotypic quantification, functional marker evaluation within the cell types identified, as well as machine-learning based topic modeling to understand cellular neighborhoods, can better explain the cellular and functional components within the intricate PDAC TME. Our analysis revealed markedly lower B and effector T cell infiltration in the treatment-naive TME of patients who eventually relapsed to date following adjuvant chemotherapy. We additionally observed the enrichment of myeloid cell populations, including tumor-associated macrophages and neutrophils, associated with PDAC tumorigenesis and worse prognosis. A deeper analysis of our functional protein expression exhibited the enrichment of GATA3 in regulatory T cells in relapsed patients, a marker which has been associated with increased accumulation and infiltration of Tregs at inflammatory sites. Our study demonstrates the importance of a targeted dissection of the spatially orchestrated immune responses and dysregulation associated with PDAC, and exemplifies how a deeper understanding of the immune and tumor components and functional aspects in situ can help improve our understanding for patients with this disease. Citation Format: Jason Yeung, Yunhao Bai, Yao Yu Yeo, Huaying Qiu, Bokai Zhu, Lara R. Heij, Enas Abu Shah, Garry P. Nolan, Shivan Sivakumar, Sizun Jiang. Spatial proteomic immune profiling of pancreatic ductal adenocarcinoma and the tumor microenvironment [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr C043.

Network-based identification of diagnosis-specific trans-omic biomarkers via integration of multiple omics data

The integration of multiple omics data promises to reveal new insights into the pathogenic mechanisms of complex human diseases, with the potential to identify avenues for the development of targeted therapies for disease subtypes. However, the extraction of diagnostic/disease-specific biomarkers from multiple omics data with biological pathway knowledge is a challenging issue in precision medicine. In this paper, we present a novel computational method to identify diagnosis-specific trans-omic biomarkers from multiple omics data. In the algorithm, we integrated multi-class sparse canonical correlation analysis (MSCCA) and molecular pathway analysis in order to derive discriminative molecular features that are correlated across different omics layers. We applied our proposed method to analyzing proteome and metabolome data of heart failure (HF), and extracted trans-omic biomarkers for HF subtypes; specifically, ischemic cardiomyopathy (ICM) and dilated cardiomyopathy (DCM). We were able to detect not only individual proteins that were previously reported from single-omics studies but also correlated protein–metabolite pairs characteristic of HF disease subtypes. For example, we identified hexokinase1(HK1)–d-fructose-6-phosphate as a paired trans-omic biomarker for DCM, which could significantly perturb amino-sugar metabolism. Our proposed method is expected to be useful for various applications in precision medicine.

MvlearnR and Shiny App for multiview learning.

The package mvlearnR and accompanying Shiny App is intended for integrating data from multiple sources or views or modalities (e.g. genomics, proteomics, clinical, and demographic data). Most existing software packages for multiview learning are decentralized and offer limited capabilities, making it difficult for users to perform comprehensive integrative analysis. The new package wraps statistical and machine learning methods and graphical tools, providing a convenient and easy data integration workflow. For users with limited programming language, we provide a Shiny Application to facilitate data integration anywhere and on any device. The methods have potential to offer deeper insights into complex disease mechanisms. mvlearnR is available from the following GitHub repository: https://github.com/lasandrall/mvlearnR. The web application is hosted on shinyapps.io and available at: https://multi-viewlearn.shinyapps.io/MultiView_Modeling/.

FORGEdb: a tool for identifying candidate functional variants and uncovering target genes and mechanisms for complex diseases.

The majority of disease-associated variants identified through genome-wide association studies are located outside of protein-coding regions. Prioritizing candidate regulatory variants and gene targets to identify potential biological mechanisms for further functional experiments can be challenging. To address this challenge, we developed FORGEdb ( https://forgedb.cancer.gov/ ; https://forge2.altiusinstitute.org/files/forgedb.html ; and https://doi.org/10.5281/zenodo.10067458 ), a standalone and web-based tool that integrates multiple datasets, delivering information on associated regulatory elements, transcription factor binding sites, and target genes for over 37 million variants. FORGEdb scores provide researchers with a quantitative assessment of the relative importance of each variant for targeted functional experiments.

Protection of p-Coumaric acid against chronic stress-induced neurobehavioral deficits in mice via activating the PKA-CREB-BDNF pathway

There is a body of evidence to suggest that chronic stress modulates neurochemical homeostasis, alters neuronal structure, inhibits neurogenesis and contributes to development of mental disorders. Chronic stress-associated mental disorders present common symptoms of cognitive impairment and depression with complex disease mechanisms. P-coumaric acid (p-CA), a natural phenolic compound, is widely distributed in vegetables, cereals and fruits. p-CA exhibits a wide range of health-related effects, including anti-oxidative-stress, anti-mutagenesis, anti-inflammation and anti-cancer activities. The current study aims to evaluate the therapeutic potential of p-CA against stress-associated mental disorders. We assessed the effect of p-CA on cognitive deficits and depression-like behavior in mice exposed to chronic restraint stress (CRS); we used network pharmacology, biochemical and molecular biological approaches to elucidate the underlying molecular mechanisms. CRS exposure caused memory impairments and depression-like behavior in mice; p-CA administration attenuated these CRS-induced memory deficits and depression-like behavior. Network pharmacology analysis demonstrated that p-CA was possibly involved in multiple targets and a variety of signaling pathways. Among them, the protein kinase A (PKA) - cAMP-response element binding protein (CREB) - brain derived neurotrophic factor (BDNF) signaling pathway was predominant and further characterized. The levels of PKA, phosphorylated CREB (pCREB) and BDNF were significantly lowered in the hippocampus of CRS mice, suggesting disruption of the PKA-CREB-BDNF signaling pathway; p-CA treatment restored the signaling pathway. Furthermore, CRS upregulated expression of proinflammatory cytokines in hippocampus, while p-CA reversed the CRS-induced effects. Our findings suggest that p-CA will offer therapeutic benefit to patients with stress-associated mental disorders.

Integrative multi-omics analysis reveals the critical role of the PBXIP1 gene in Alzheimer's disease.

Alzheimer's disease (AD) is a neurodegenerative disorder, and its strongest risk factor is aging. A few studies have explored the relationship between aging and AD, while the underlying mechanism remains unclear. We assembled data across multi-omics (i.e., epigenetics, transcriptomics, and proteomics, based on frozen tissues from the dorsolateral prefrontal cortex) and neuropathological and clinical traits from the Religious Orders Study and Rush Memory and Aging Project (ROSMAP). Aging was assessed using six DNA methylation clocks (including the Horvath clock, Hannum clock, Levine clock, HorvathSkin clock, Lin clock, and Cortical clock) that capture mortality risk in literature. After accounting for age, we first identified a gene module (including 263 genes) that was related to the integrated aging measure of six clocks, as well as three neuropathological traits of AD (i.e., β-amyloid, Tau tangles, and tangle density). Interestingly, among 20 key genes with top intramodular connectivity of the module, PBXIP1 was the only one that was significantly associated with all three neuropathological traits of AD at the protein level after Bonferroni correction. Furthermore, PBXIP1 was associated with the clinical diagnosis of AD in both ROSMAP and three independent datasets. Moreover, PBXIP1 may be related to AD through its role in astrocytes and hippocampal neurons, and the mTOR pathway. The results suggest the critical role of PBXIP1 in AD and support the potential and feasibility of using multi-omics data to investigate mechanisms of complex diseases. However, more validations in different populations and experiments invitro and invivo are required in the future.

Rapid and signal crowdedness-robust in situ sequencing through hybrid block coding.

Spatial transcriptomics technology has revolutionized our understanding of cell types and tissue organization, opening possibilities for researchers to explore transcript distributions at subcellular levels. However, existing methods have limitations in resolution, sensitivity, or speed. To overcome these challenges, we introduce SPRINTseq (Spatially Resolved and signal-diluted Next-generation Targeted sequencing), an innovative in situ sequencing strategy that combines hybrid block coding and molecular dilution strategies. Our method enables fast and sensitive high-resolution data acquisition, as demonstrated by recovering over 142 million transcripts using a 108-gene panel from 453,843 cells from four mouse brain coronal slices in less than 2 d. Using this advanced technology, we uncover the cellular and subcellular molecular architecture of Alzheimer's disease, providing additional information into abnormal cellular behaviors and their subcellular mRNA distribution. This improved spatial transcriptomics technology holds great promise for exploring complex biological processes and disease mechanisms.

A model-based clustering via mixture of hierarchical models with covariate adjustment for detecting differentially expressed genes from paired design

The causes of many complex human diseases are still largely unknown. Genetics plays an important role in uncovering the molecular mechanisms of complex human diseases. A key step to characterize the genetics of a complex human disease is to unbiasedly identify disease-associated gene transcripts on a whole-genome scale. Confounding factors could cause false positives. Paired design, such as measuring gene expression before and after treatment for the same subject, can reduce the effect of known confounding factors. However, not all known confounding factors can be controlled in a paired/match design. Model-based clustering, such as mixtures of hierarchical models, has been proposed to detect gene transcripts differentially expressed between paired samples. To the best of our knowledge, no model-based gene clustering methods have the capacity to adjust for the effects of covariates yet. In this article, we proposed a novel mixture of hierarchical models with covariate adjustment in identifying differentially expressed transcripts using high-throughput whole-genome data from paired design. Both simulation study and real data analysis show the good performance of the proposed method.

A Gene Set-Integrated Approach for Predicting Disease-Associated Genes.

It is important to identify disease-associated genes for studying the pathogenic mechanism of complex diseases. Recently, models for disease gene prediction are dominantly based on molecular expression data and networks, including gene expression, protein expression, co-expression networks, protein-protein interaction networks, etc. One limitation of these methods is that they do not consider the knowledge of annotated gene sets representing known pathways or functionally-related sets of genes. In this study, we propose a new approach to predict disease-associated genes by integrating annotated gene sets data from the Molecular Signature Database (MSigDB). It first represents and integrates the different types of annotated gene sets in the MSigDB database in the form of the signal matrix. It then uses the signal matrix as the gene feature to train the disease gene prediction model. We compare our method with existing methods in predicting genes for five complex diseases. The results show that our method is superior to other methods. Further, we perform a case study on autism spectrum disorder (ASD). We find that ASD predictions are associated with ASD based on the statistical analysis of biological networks and independent ASD studies. The source code, prediction results and datasets are publicly available on https://github.com/genemine/GSI.git.

Prediction of circRNA-MiRNA Association Using Singular Value Decomposition and Graph Neural Networks.

A large number of experimental studies have shown that circRNAs can act as molecular sponges of microRNAs, interacting with miRNAs to regulate gene expression levels, thereby affecting the development of human diseases. Exploring the potential associations between circRNAs and miRNAs can help understand complex disease mechanisms. Considering that biological experiments are time-consuming and labor-intensive, this study proposes a computational model using a graph neural network and singular value decomposition (CMASG) for circRNA-miRNA association prediction. Specifically, graph neural networks are used to learn nonlinear feature representations of nodes, followed by matrix factorization algorithms to learn linear feature representations of nodes, and then combined feature representations learned from different perspectives. Finally, the lightGBM algorithm was used for circRNA-miRNA association prediction. The proposed CMASG model achieved an AUC value of 0.8804. The experimental results demonstrate the superiority and effectiveness of the CMASG model in predicting circRNA-miRNA association tasks.