Investigating the association of CD36 gene polymorphisms (rs1761667 and rs1527483) with T2DM and dyslipidemia: Statistical analysis, machine learning based prediction, and meta-analysis

CD36 (cluster of differentiation 36) is a membrane protein involved in lipid metabolism and has been linked to pathological conditions associated with metabolic disorders, such as diabetes and dyslipidemia. A case-control study was conducted and included 177 patients with type-2 diabetes mellitus (T2DM) and 173 control subjects to study the involvement of CD36 gene rs1761667 (G>A) and rs1527483 (C>T) polymorphisms in the pathogenesis of T2DM and dyslipidemia among Jordanian population. Lipid profile, blood sugar, gender and age were measured and recorded. Also, genotyping analysis for both polymorphisms was performed. Following statistical analysis, 10 different neural networks and machine learning (ML) tools were used to predict subjects with diabetes or dyslipidemia. Towards further understanding of the role of CD36 protein and gene in T2DM and dyslipidemia, a protein-protein interaction network and meta-analysis were carried out. For both polymorphisms, the genotypic frequencies were not significantly different between the two groups (p > 0.05). On the other hand, some ML tools like multilayer perceptron gave high prediction accuracy (≥ 0.75) and Cohen’s kappa (κ) (≥ 0.5). Interestingly, in K-star tool, the accuracy and Cohen’s κ values were enhanced by including the genotyping results as inputs (0.73 and 0.46, respectively, compared to 0.67 and 0.34 without including them). This study confirmed, for the first time, that there is no association between CD36 polymorphisms and T2DM or dyslipidemia among Jordanian population. Prediction of T2DM and dyslipidemia, using these extensive ML tools and based on such input data, is a promising approach for developing diagnostic and prognostic prediction models for a wide spectrum of diseases, especially based on large medical databases.

Download Full-text

Comparison of Multivariable Logistic Regression and Other Machine Learning Algorithms for Prognostic Prediction Studies in Pregnancy Care: Systematic Review and Meta-Analysis (Preprint)

10.2196/preprints.16503 ◽

2019 ◽

Author(s):

Herdiantri Sufriyana ◽

Atina Husnayain ◽

Ya-Lin Chen ◽

Chao-Yang Kuo ◽

Onkar Singh ◽

...

Keyword(s):

Machine Learning ◽

Systematic Review ◽

Pregnancy Outcomes ◽

Prediction Models ◽

Meta Analysis ◽

Machine Learning Algorithms ◽

Gradient Boosting ◽

Pregnancy Care ◽

Prognostic Prediction ◽

In Pregnancy

BACKGROUND Predictions in pregnancy care are complex because of interactions among multiple factors. Hence, pregnancy outcomes are not easily predicted by a single predictor using only one algorithm or modeling method. OBJECTIVE This study aims to review and compare the predictive performances between logistic regression (LR) and other machine learning algorithms for developing or validating a multivariable prognostic prediction model for pregnancy care to inform clinicians’ decision making. METHODS Research articles from MEDLINE, Scopus, Web of Science, and Google Scholar were reviewed following several guidelines for a prognostic prediction study, including a risk of bias (ROB) assessment. We report the results based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies were primarily framed as PICOTS (population, index, comparator, outcomes, timing, and setting): Population: men or women in procreative management, pregnant women, and fetuses or newborns; Index: multivariable prognostic prediction models using non-LR algorithms for risk classification to inform clinicians’ decision making; Comparator: the models applying an LR; Outcomes: pregnancy-related outcomes of procreation or pregnancy outcomes for pregnant women and fetuses or newborns; Timing: pre-, inter-, and peripregnancy periods (predictors), at the pregnancy, delivery, and either puerperal or neonatal period (outcome), and either short- or long-term prognoses (time interval); and Setting: primary care or hospital. The results were synthesized by reporting study characteristics and ROBs and by random effects modeling of the difference of the logit area under the receiver operating characteristic curve of each non-LR model compared with the LR model for the same pregnancy outcomes. We also reported between-study heterogeneity by using τ2 and I2. RESULTS Of the 2093 records, we included 142 studies for the systematic review and 62 studies for a meta-analysis. Most prediction models used LR (92/142, 64.8%) and artificial neural networks (20/142, 14.1%) among non-LR algorithms. Only 16.9% (24/142) of studies had a low ROB. A total of 2 non-LR algorithms from low ROB studies significantly outperformed LR. The first algorithm was a random forest for preterm delivery (logit AUROC 2.51, 95% CI 1.49-3.53; I2=86%; τ2=0.77) and pre-eclampsia (logit AUROC 1.2, 95% CI 0.72-1.67; I2=75%; τ2=0.09). The second algorithm was gradient boosting for cesarean section (logit AUROC 2.26, 95% CI 1.39-3.13; I2=75%; τ2=0.43) and gestational diabetes (logit AUROC 1.03, 95% CI 0.69-1.37; I2=83%; τ2=0.07). CONCLUSIONS Prediction models with the best performances across studies were not necessarily those that used LR but also used random forest and gradient boosting that also performed well. We recommend a reanalysis of existing LR models for several pregnancy outcomes by comparing them with those algorithms that apply standard guidelines. CLINICALTRIAL PROSPERO (International Prospective Register of Systematic Reviews) CRD42019136106; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=136106

Download Full-text

Comparison of Multivariable Logistic Regression and Other Machine Learning Algorithms for Prognostic Prediction Studies in Pregnancy Care: Systematic Review and Meta-Analysis

JMIR Medical Informatics ◽

10.2196/16503 ◽

2020 ◽

Vol 8 (11) ◽

pp. e16503

Author(s):

Herdiantri Sufriyana ◽

Atina Husnayain ◽

Ya-Lin Chen ◽

Chao-Yang Kuo ◽

Onkar Singh ◽

...

Keyword(s):

Machine Learning ◽

Systematic Review ◽

Pregnancy Outcomes ◽

Prediction Models ◽

Meta Analysis ◽

Machine Learning Algorithms ◽

Gradient Boosting ◽

Pregnancy Care ◽

Prognostic Prediction ◽

In Pregnancy

Background Predictions in pregnancy care are complex because of interactions among multiple factors. Hence, pregnancy outcomes are not easily predicted by a single predictor using only one algorithm or modeling method. Objective This study aims to review and compare the predictive performances between logistic regression (LR) and other machine learning algorithms for developing or validating a multivariable prognostic prediction model for pregnancy care to inform clinicians’ decision making. Methods Research articles from MEDLINE, Scopus, Web of Science, and Google Scholar were reviewed following several guidelines for a prognostic prediction study, including a risk of bias (ROB) assessment. We report the results based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Studies were primarily framed as PICOTS (population, index, comparator, outcomes, timing, and setting): Population: men or women in procreative management, pregnant women, and fetuses or newborns; Index: multivariable prognostic prediction models using non-LR algorithms for risk classification to inform clinicians’ decision making; Comparator: the models applying an LR; Outcomes: pregnancy-related outcomes of procreation or pregnancy outcomes for pregnant women and fetuses or newborns; Timing: pre-, inter-, and peripregnancy periods (predictors), at the pregnancy, delivery, and either puerperal or neonatal period (outcome), and either short- or long-term prognoses (time interval); and Setting: primary care or hospital. The results were synthesized by reporting study characteristics and ROBs and by random effects modeling of the difference of the logit area under the receiver operating characteristic curve of each non-LR model compared with the LR model for the same pregnancy outcomes. We also reported between-study heterogeneity by using τ2 and I2. Results Of the 2093 records, we included 142 studies for the systematic review and 62 studies for a meta-analysis. Most prediction models used LR (92/142, 64.8%) and artificial neural networks (20/142, 14.1%) among non-LR algorithms. Only 16.9% (24/142) of studies had a low ROB. A total of 2 non-LR algorithms from low ROB studies significantly outperformed LR. The first algorithm was a random forest for preterm delivery (logit AUROC 2.51, 95% CI 1.49-3.53; I2=86%; τ2=0.77) and pre-eclampsia (logit AUROC 1.2, 95% CI 0.72-1.67; I2=75%; τ2=0.09). The second algorithm was gradient boosting for cesarean section (logit AUROC 2.26, 95% CI 1.39-3.13; I2=75%; τ2=0.43) and gestational diabetes (logit AUROC 1.03, 95% CI 0.69-1.37; I2=83%; τ2=0.07). Conclusions Prediction models with the best performances across studies were not necessarily those that used LR but also used random forest and gradient boosting that also performed well. We recommend a reanalysis of existing LR models for several pregnancy outcomes by comparing them with those algorithms that apply standard guidelines. Trial Registration PROSPERO (International Prospective Register of Systematic Reviews) CRD42019136106; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=136106

Download Full-text

MicroRNA-based regulation of genomics and transcriptomics of inflammatory cytokines in COVID-19

10.1101/2021.06.08.21258565 ◽

2021 ◽

Author(s):

Manoj Khokhar ◽

Sojit Tomo ◽

Purvi Purohit

Keyword(s):

Transcription Factors ◽

Inflammatory Response ◽

In Silico ◽

Wide Spectrum ◽

Interaction Network ◽

In Silico Analysis ◽

Enrichment Analysis ◽

Microarray Dataset ◽

Therapeutic Strategies ◽

Protein Protein Interaction

Background: Coronavirus disease 2019 is characterized by the elevation of a wide spectrum of inflammatory mediators which are associated with poor disease outcomes. We aimed at an in-silico analysis of regulatory microRNA and their transcription factors (TF) for these inflammatory genes that may help to devise potential therapeutic strategies in the future. Methods: The cytokine regulating immune-expressed genes (CRIEG) were sorted from literature and GEO microarray dataset and their co-differentially expressed miRNA and transcription factors were predicted from publicly available databases. Enrichment analysis was done through mienturnet, MiEAA, and Gene Ontology, and pathways predicted by KEGG and Reactome pathways. The functional and regulatory features were analyzed and visualized through Cytoscape. Results: Sixteen CRIEG were observed to have a significant protein-protein interaction network. The ontological analysis revealed significantly enriched pathways for biological processes, molecular functions, and cellular components. The search performed in the miRNA database yielded 10 miRNAs that are significantly involved in the regulation of these genes and their transcription factors. Conclusion: An In-Silico representation of a network involving miRNAs, CRIEGs, and TF which take part in the inflammatory response in COVID-19 has been elucidated. These regulatory factors may have potentially critical roles in the inflammatory response in COVID-19 and may be explored further for the development of targeted therapeutic strategies and mechanistic validation.

Download Full-text

Prognostic prediction models for pregnancy complications in women with gestational diabetes: a protocol for systematic review, critical appraisal and meta-analysis

Systematic Reviews ◽

10.1186/s13643-019-1151-0 ◽

2019 ◽

Vol 8 (1) ◽

Cited By ~ 7

Author(s):

Shamil D. Cooray ◽

Jacqueline A. Boyle ◽

Georgia Soldatos ◽

Lihini A. Wijeyaratne ◽

Helena J. Teede

Keyword(s):

Systematic Review ◽

Decision Making ◽

Gestational Diabetes ◽

Pregnancy Complications ◽

Prediction Models ◽

Absolute Risk ◽

Meta Analysis ◽

External Validation ◽

The Absolute ◽

Prognostic Prediction

Abstract Background Gestational diabetes (GDM) is increasingly common and has significant implications during pregnancy and for the long-term health of the mother and offspring. However, it is a heterogeneous condition with inter-related factors including ethnicity, body mass index and gestational weight gain significantly modifying the absolute risk of complications at an individual level. Predicting the risk of pregnancy complications for an individual woman with GDM presents a useful adjunct to therapeutic decision-making and patient education. Diagnostic prediction models for GDM are prevalent. In contrast, prediction models for risk of complications in those with GDM are relatively novel. This study will systematically review published prognostic prediction models for pregnancy complications in women with GDM, describe their characteristics, compare performance and assess methodological quality and applicability. Methods Studies will be identified by searching MEDLINE and Embase electronic databases. Title and abstract screening, full-text review and data extraction will be completed independently by two reviewers. The included studies will be systematically assessed for risk of bias and applicability using appropriate tools designed for prediction modelling studies. Extracted data will be tabulated to facilitate qualitative comparison of published prediction models. Quantitative data on predictive performance of these models will be synthesised with meta-analyses if appropriate. Discussion This review will identify and summarise all published prognostic prediction models for pregnancy complications in women with GDM. We will compare model performance across different settings and populations with meta-analysis if appropriate. This work will guide subsequent phases in the prognosis research framework: further model development, external validation and model updating, and impact assessment. The ultimate model will estimate the absolute risk of pregnancy complications for women with GDM and will be implemented into routine care as an evidence-based GDM complication risk prediction model. It is anticipated to offer value to women and their clinicians with individualised risk assessment and may assist decision-making. Ultimately, this systematic review is an important step towards a personalised risk-stratified model-of-care for GDM to allow preventative and therapeutic interventions for the maximal benefit to women and their offspring, whilst sparing expense and harm for those at low risk. Systematic review registration PROSPERO registration number CRD42019115223

Download Full-text

Machine Learning Prediction Models for Gestational Diabetes Mellitus: A meta- analysis (Preprint)

Journal of Medical Internet Research ◽

10.2196/26634 ◽

2020 ◽

Author(s):

Zheqing Zhang ◽

Luqian Yang ◽

Wentao Han ◽

Yaoyu Wu ◽

Linhui Zhang ◽

...

Keyword(s):

Diabetes Mellitus ◽

Machine Learning ◽

Gestational Diabetes ◽

Gestational Diabetes Mellitus ◽

Prediction Models ◽

Meta Analysis

Download Full-text

MiR-574-5P, miR-1827, And miR-4429 As Potential Biomarkers For Schizophrenia

10.21203/rs.3.rs-844121/v1 ◽

2021 ◽

Author(s):

Omran Davarinejad ◽

Sajad Najafi ◽

Hossein Zhaleh ◽

Farzaneh Golmohammadi ◽

Farnaz Radmehr ◽

...

Keyword(s):

Mrna Expression ◽

Meta Analysis ◽

Interaction Network ◽

Receiver Operating Curve ◽

Healthy Controls ◽

Protein Protein Interaction ◽

Mirna Array ◽

Expression Arrays ◽

Limma Package ◽

Novel Biomarkers

Abstract Schizophrenia is a severe chronic debilitating disorder with millions of affected individuals. Lack of a reliable mollecular diagnostic invokes the identification of novel biomarkers. To elucidate the molecular basis of the disease, two mRNA expression arrays including GSE93987 and GSE38485, and one miRNA array, GSE54914, were downloaded from GEO, and meta-analysis was performed for mRNA expression arrays by employment of metaDE package. By WGCNA package, we performed network analysis for both mRNA expression arrays separately. Then, we made protein-protein interaction network for significant modules. Limma package was employed to analyze the miRNA array and dysregulated miRNAs (DEMs) were identified. Using genes of significant modules and DEMs, a mRNA-miRNA network was constructed and hub genes and miRNAs were identified. To confirm the dysregulation of genes, expression values were evaluated by available datasets including GEO series GSE62333, GSE93987, and GSE38485. The ability of the detected hub miRNAs to discriminate Schizophrenia from healthy controls was evaluated by assessing the receiver-operating curve. Finally, by performing Real-Time PCR, the expression level of genes and miRNAs were evaluated in 40 Schizophrenia patients compared with healthy controls. The results confirmed dysregulation of hsa-miR-574-5P, hsa-miR-1827, hsa-miR-4429, CREBRF, ARPP19, TGFBR2, and YWHAZ in blood samples of schizophrenia patients.

Download Full-text

Supervised machine learning models and protein-protein interaction network analysis of gene expression profiles induced by omega-3 polyunsaturated fatty acids

Current Chinese Science ◽

10.2174/2210298102666220112114505 ◽

2022 ◽

Vol 02 ◽

Author(s):

Sergey Shityakov ◽

Jane Pei-Chen Chang ◽

Ching-Fang Sun ◽

David Ta-Wei Guu ◽

Thomas Dandekar ◽

...

Keyword(s):

Gene Expression ◽

Machine Learning ◽

Protein Interaction ◽

Expression Profiles ◽

Gene Expression Profiles ◽

Interaction Network ◽

Supervised Machine Learning ◽

Omega 3 ◽

Protein Protein Interaction ◽

Protein Protein Interaction Network

Background: Omega-3 polyunsaturated fatty acids (PUFAs), such as eicosapentaenoic (EPA) and docosahexaenoic (DHA) acids, have beneficial effects on human health, but their effect on gene expression in elderly individuals (age ≥ 65) is largely unknown. In order to examine this, the gene expression profiles were analyzed in the healthy subjects (n = 96) at baseline and after 26 weeks of supplementation with EPA+DHA to determine up-regulated and down-regulated dif-ferentially expressed genes (DEGs) triggered by PUFAs. The protein-protein interaction (PPI) networks were constructed by mapping these DEGs to a human interactome and linking them to the specific pathways. Objective: This study aimed to implement supervised machine learning models and protein-protein interaction network analysis of gene expression profiles induced by PUFAs. Methods: The transcriptional profile of GSE12375 was obtained from the Gene Expression Om-nibus database, which is based on the Affymetrix NuGO array. The probe cell intensity data were converted into the gene expression values, and the background correction was performed by the multi-array average algorithm. The LIMMA (Linear Models for Microarray Data) algo-rithm was implemented to identify relevant DEGs at baseline and after 26 weeks of supplemen-tation with a p-value < 0.05. The DAVID web server was used to identify and construct the en-riched KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways. Finally, the construction of machine learning (ML) models, including logistic regression, naïve Bayes, and deep neural networks, were implemented for the analyzed DEGs associated with the specific pathways. Results: The results revealed that up-regulated DEGs were associated with neurotrophin/MAPK signaling, whereas the down-regulated DEGs were linked to cancer, acute myeloid leukemia, and long-term depression pathways. Additionally, ML approaches were able to cluster the EPA/DHA-treated and control groups by the logistic regression performing the best. Conclusion: Overall, this study highlights the pivotal changes in DEGs induced by PUFAs and provides the rationale for the implementation of ML algorithms as predictive models for this type of biomedical data.

Download Full-text

Integrating protein networks and machine learning for disease stratification in the Hereditary Spastic Paraplegias

10.1101/2021.01.14.425874 ◽

2021 ◽

Author(s):

Nikoleta Vavouraki ◽

James E. Tomkins ◽

Eleanna Kara ◽

Henry Houlden ◽

John Hardy ◽

...

Keyword(s):

Machine Learning ◽

Protein Interaction ◽

Protein Interaction Network ◽

Molecular Mechanisms ◽

Topological Analysis ◽

Interaction Network ◽

Protein Protein Interaction ◽

Hereditary Spastic Paraplegias ◽

Core Proteins ◽

Protein Protein Interaction Network

AbstractThe Hereditary Spastic Paraplegias are a group of neurodegenerative diseases characterized by spasticity and weakness in the lower body. Despite the identification of causative mutations in over 70 genes, the molecular aetiology remains unclear. Due to the combination of genetic diversity and variable clinical presentation, the Hereditary Spastic Paraplegias are a strong candidate for protein-protein interaction network analysis as a tool to understand disease mechanism(s) and to aid functional stratification of phenotypes. In this study, experimentally validated human protein-protein interactions were used to create a protein-protein interaction network based on the causative Hereditary Spastic Paraplegia genes. Network evaluation as a combination of both topological analysis and functional annotation led to the identification of core proteins in putative shared biological processes such as intracellular transport and vesicle trafficking. The application of machine learning techniques suggested a functional dichotomy linked with distinct sets of clinical presentations, suggesting there is scope to further classify conditions currently described under the same umbrella term of Hereditary Spastic Paraplegias based on specific molecular mechanisms of disease.

Download Full-text

FYN and CD247: key Genes for Septic Shock Based on Bioinformatics and Meta-Analysis

Combinatorial Chemistry & High Throughput Screening ◽

10.2174/1386207324666210816123508 ◽

2021 ◽

Vol 24 ◽

Author(s):

Yue Jiang ◽

Qian Miao ◽

Lin Hu ◽

Tingyan Zhou ◽

Yingchun Hu ◽

...

Keyword(s):

Gene Expression ◽

Septic Shock ◽

Survival Analysis ◽

Expression Profiles ◽

Meta Analysis ◽

Interaction Network ◽

Protein Protein Interaction ◽

Key Genes ◽

Protein Protein Interaction Network ◽

Geo Database

Background: Septic shock is sepsis accompanied by hemodynamic instability and high clinical mortality. Material and Methods: GSE95233, GSE57065, GSE131761 gene-expression profiles of healthy control subjects and septic shock patients were downloaded from the Gene-Expression Omnibus (GEO) database, and differences of expression profiles and their intersection were analysed using GEO2R. Function and pathway enrichment analysis was performed on common differentially expressed genes (DEG), and key genes for septic shock were screened using a protein-protein interaction network created with STRING. Also, data from the GEO database were used for survival analysis for key genes, and a meta-analysis was used to explore expression trends of core genes. Finally, high-throughput sequencing using the blood of a murine sepsis model was performed to analyse the expression of CD247 and FYN in mice. Results: A total of 539 DEGs were obtained (p < 0.05). Gene ontology analysis showed that key genes were enriched in functions, such as immune response and T cell activity, and DEGs were enriched in signal pathways, such as T cell receptors. FYN and CD247 are in the centre of the protein-protein interaction network, and survival analysis found that they are positively correlated with survival from sepsis. Further, meta-analysis results showed that FYN could be useful for the prognosis of patients, and CD247 might distinguish between sepsis and systemic inflammatory response syndrome patients. Finally, RNA sequencing using a mouse septic shock model showed low expression of CD247 and FYN in this model. Conclusion: FYN and CD247 are expected to become new biomarkers of septic shock.

Download Full-text