Analysis of microRNA expression using machine learning

Diagnostic potential of miR-200 family members in gingival crevicular fluid for chronic periodontitis: correlation with clinical parameters and therapeutic implications

OBJECTIVE

The purpose of this study was to evaluate the potential of miR-200 family members in gingival crevicular fluid (GCF) as diagnostic biomarkers for chronic periodontitis (CP), aiming to provide valuable insights for the early detection and management of the disease.

METHODS

GSE89081 dataset profiled miRNAs in GCF derived from 5 healthy and 5 periodontitis was analyzed by GEO2R. Quantitative real-time PCR was used to quantify the expression levels of miR-200 family members (miR-200a-3p, miR-200a-5p, miR-200b-3p, miR-200b-5p, miR-200c-3p, miR-200c-5p, miR-141-3p, miR-141-5p, and miR-429) in the GCF samples from 103 CP patients and 113 healthy controls. Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic potential of miR-200 family members in differentiating CP patients from healthy controls.

RESULTS

By analyzing the GSE89081 dataset, miR-200a-5p, miR-200b-5p and miR-200c-5p were significantly upregulated in GCF of the CP patients compared to the healthy control. In this study, miR-200a-3p, miR-200a-5p, miR-200b-3p, miR-200b-5p, miR-200c-3p, miR-200c-5p were significantly increased in GCF of CP patients compared to the healthy control, while miR-141 and miR-429 did not show significant differences. MiR-200a, -200b and 200c had good diagnostic value, and when these miRNAs were combined, they demonstrated excellent diagnostic value for CP with an AUC of 0.997, sensitivity of 99.03%, and specificity of 98.23%. MiR-200a, -200b and 200c in GCF showed significant and positive correlation with plaque index (PI), gingival index (GI), bleeding on probing (BOP), clinical attachment level (CAL), and probing pocket depth (PPD).

CONCLUSION

MiR-200a, -200b and 200c in GCF may serve as potential biomarkers for the early diagnosis of CP, which was correlated with clinical parameters, being therapeutic targets for CP.

Deciphering the Transcriptomic Heterogeneity of Duodenal Coeliac Disease Biopsies

Coeliac disease (CD) is a clinically heterogeneous autoimmune disease with variable presentation and progression triggered by gluten intake. Molecular or genetic factors contribute to disease heterogeneity, but the reasons for different outcomes are poorly understood. Transcriptome studies of tissue biopsies from CD patients are scarce. Here, we present a high-resolution analysis of the transcriptomes extracted from duodenal biopsies of 24 children and adolescents with active CD and 21 individuals without CD but with intestinal afflictions as controls. The transcriptomes of CD patients divide into three groups-a mixed group presenting the control cases, and CD-low and CD-high groups referring to lower and higher levels of CD severity. Persistence of symptoms was weakly associated with subgroup, but the highest marsh stages were present in subgroup CD-high, together with the highest cell cycle rates as an indicator of virtually complete villous atrophy. Considerable variation in inflammation-level between subgroups was further deciphered into immune cell types using cell type de-convolution. Self-organizing maps portrayal was applied to provide high-resolution landscapes of the CD-transcriptome. We find asymmetric patterns of miRNA and long non-coding RNA and discuss the effect of epigenetic regulation. Expression of genes involved in interferon gamma signaling represent suitable markers to distinguish CD from non-CD cases. Multiple pathways overlay in CD biopsies in different ways, giving rise to heterogeneous transcriptional patterns, which potentially provide information about etiology and the course of the disease.

MicroRNA-34 family in breast cancer: from research to therapeutic potential

MicroRNA (miRNA)-34 family (miR-34s), including miR-34a/b/c, is the most well studied non-coding RNAs that regulate gene expression post-transcriptionally. The miR-34s mediates the tumor suppressor function of p53 in the pathogenesis of breast cancer by targeting different oncogenes. This review focuses on the anti-oncogenic regulation of the miR-34s, emphasizing the major signaling pathways that are involved in the modulation of miR-34s in breast cancer. Moreover, it highlights how epigenetic modification by the p53/miR-34s axis regulates the proliferation, invasiveness, chemoresistance, and sternness of breast cancer. A better understanding of the molecular mechanisms of miR-34s will open new opportunities for the development of novel therapeutic strategies and define a new approach in identifying potential biomarkers for early diagnosis of breast cancer.

Cartography of Pathway Signal Perturbations Identifies Distinct Molecular Pathomechanisms in Malignant and Chronic Lung Diseases

Lung diseases are described by a wide variety of developmental mechanisms and clinical manifestations. Accurate classification and diagnosis of lung diseases are the bases for development of effective treatments. While extensive studies are conducted toward characterization of various lung diseases at molecular level, no systematic approach has been developed so far. Here we have applied a methodology for pathway-centered mining of high throughput gene expression data to describe a wide range of lung diseases in the light of shared and specific pathway activity profiles. We have applied an algorithm combining a Pathway Signal Flow (PSF) algorithm for estimation of pathway activity deregulation states in lung diseases and malignancies, and a Self Organizing Maps algorithm for classification and clustering of the pathway activity profiles. The analysis results allowed clearly distinguish between cancer and non-cancer lung diseases. Lung cancers were characterized by pathways implicated in cell proliferation, metabolism, while non-malignant lung diseases were characterized by deregulations in pathways involved in immune/inflammatory response and fibrotic tissue remodeling. In contrast to lung malignancies, chronic lung diseases had relatively heterogeneous pathway deregulation profiles. We identified three groups of interstitial lung diseases and showed that the development of characteristic pathological processes, such as fibrosis, can be initiated by deregulations in different signaling pathways. In conclusion, this paper describes the pathobiology of lung diseases from systems viewpoint using pathway centered high-dimensional data mining approach. Our results contribute largely to current understanding of pathological events in lung cancers and non-malignant lung diseases. Moreover, this paper provides new insight into molecular mechanisms of a number of interstitial lung diseases that have been studied to a lesser extent.

Revisiting Chaos Theorem to Understand the Nature of miRNAs in Response to Drugs of Abuse

Just like Matryoshka dolls, biological systems follow a hierarchical order that is based on dynamic bidirectional communication among its components. In addition to the convoluted inter-relationships, the complexity of each component spans several folds. Therefore, it becomes rather challenging to investigate phenotypes resulting from these networks as it requires the integration of reductionistic and holistic approaches. One dynamic system is the transcriptome which comprises a variety of RNA species. Some, like microRNAs, have recently received a lot of attention. miRNAs are very pleiotropic and have been considered as therapeutic and diagnostic candidates in the biomedical fields. In this review, we survey miRNA profiles in response to drugs of abuse (DA) using 118 studies. After providing a summary of miRNAs related to substance use disorders (SUD), general patterns of miRNA signatures are compared among studies for single or multiple drugs of abuse. Then, current challenges and drawbacks in the field are discussed. Finally, we provide support for considering miRNAs as a chaotic system in normal versus disrupted states particularly in SUD and propose an integrative approach for studying and analyzing miRNA data.

Personalized Disease Phenotypes from Massive OMICs Data

Application of new high-throughput technologies in molecular medicine collects massive data for hundreds to thousands of persons in large cohort studies by characterizing the phenotype of each individual on a personalized basis. The chapter aims at increasing our understanding of disease genesis and progression and to improve diagnosis and treatment. New methods are needed to handle such “big data.” Machine learning enables one to recognize and to visualize complex data patterns and to make decisions potentially relevant for diagnosis and treatment. The authors address these tasks by applying the method of self-organizing maps and present worked examples from different disease entities of the colon ranging from inflammation to cancer.

Portraying the Expression Landscapes of B-CellLymphoma-Intuitive Detection of Outlier Samples and of Molecular Subtypes

We present an analytic framework based on Self-Organizing Map (SOM) machine learning to study large scale patient data sets. The potency of the approach is demonstrated in a case study using gene expression data of more than 200 mature aggressive B-cell lymphoma patients. The method portrays each sample with individual resolution, characterizes the subtypes, disentangles the expression patterns into distinct modules, extracts their functional context using enrichment techniques and enables investigation of the similarity relations between the samples. The method also allows to detect and to correct outliers caused by contaminations. Based on our analysis, we propose a refined classification of B-cell Lymphoma into four molecular subtypes which are characterized by differential functional and clinical characteristics.