Data integration by multi-tuning parameter elastic net regression

Exploration of Variables Predicting Sense of School Belonging Using the Machine Learning Method-Group Mnet

The purpose of this study was to explore variables related to school belonging from a holistic perspective, including a large number of variables in one model, different to the traditional analytical method. Using 2015 data from the Program for International Student Assessment (PISA), we sought to identify variables related to school belonging by searching for hundreds of predictors in one model using the group Mnet machine learning technique. The study repeated 100 rounds of model building after random data splitting. After exploring 504 variables (384 student and 99 parent), 32 variables were finally selected after selection counts. Variables predicting a sense of school belonging were categorized as individual/parent variables (e.g. motivation to achieve, tendency to cooperative learning, parental support) and school-related variables (e.g. school satisfaction, peer/teacher relationship, learning/physical activities). The significance and implications of the study as well as future research topics were discussed.

Spatial and temporal resolution of metabolic dysregulation in the Sugen hypoxia model of pulmonary hypertension

Although PAH is partially attributed to disordered metabolism, previous human studies have mostly examined circulating metabolites at a single time point, potentially overlooking crucial disease biology. Current knowledge gaps include an understanding of temporal changes that occur within and across relevant tissues, and whether observed metabolic changes might contribute to disease pathobiology. We utilized targeted tissue metabolomics in the Sugen hypoxia (SuHx) rodent model to investigate tissue-specific metabolic relationships with pulmonary hypertensive features over time using regression modeling and time-series analysis. Our hypotheses were that some metabolic changes would precede phenotypic changes, and that examining metabolic interactions across heart, lung, and liver tissues would yield insight into interconnected metabolic mechanisms. To support the relevance of our findings, we sought to establish links between SuHx tissue metabolomics and human PAH -omics data using bioinformatic predictions. Metabolic differences between and within tissue types were evident by Day 7 postinduction, demonstrating distinct tissue-specific metabolism in experimental pulmonary hypertension. Various metabolites demonstrated significant tissue-specific associations with hemodynamics and RV remodeling. Individual metabolite profiles were dynamic, and some metabolic shifts temporally preceded the emergence of overt pulmonary hypertension and RV remodeling. Metabolic interactions were observed such that abundance of several liver metabolites modulated lung and RV metabolite-phenotype relationships. Taken all together, regression analyses, pathway analyses and time-series analyses implicated aspartate and glutamate signaling and transport, glycine homeostasis, lung nucleotide abundance, and oxidative stress as relevant to early PAH pathobiology. These findings offer valuable insights into potential targets for early intervention in PAH.

Which COVID-19 information really impacts stock markets?

Information about the COVID-19 pandemic abounds, but which COVID-19 data actually impacts stock prices? We investigate which measures of COVID-19 matter most by applying elastic net regression for measure selection using a sample of the 35 largest stock markets. Out of 24 measures, COVID-19 related Google search trends, the stringency of government responses and media hype prevail during the height of the COVID-19 crisis. These measures proxy for COVID-19 related uncertainty, the economic impact of lockdowns and panic-driven media attention, respectively, summarizing key aspects of COVID-19 that move stock markets. Moreover, geographical proximity to the virus’s outbreak and a country’s development level also matter in terms of impact.

Correlating metal exposures and dietary habits with hyperuricemia in a large urban elderly cohort by artificial intelligence

Epidemiological studies using conventional statistical methods have reported an association between individual metal exposure and hyperuricemia (HUA). There is also evidence that diet may influence HUA development, although the available data are inconsistent. We therefore used an elastic net regression (ENR) model to screen the usefulness of various environmental and dietary factors as predictors of HUA in a large sample cohort. This study included 6217 subjects drawn from the Shenzhen Aging Related Disorder Cohort. We obtained information on the subjects' dietary habits via face-to-face interviews and used inductively coupled plasma mass spectrometry (ICP-MS) to measure the urinary concentrations of 24 metals to which elderly persons in large urban areas may be exposed. An elastic net regression (ENR) model was generated to screen the utility of the metals and dietary factors as predictors of HUA, and we demonstrated the superiority of the ENR model by comparing it to a traditional logistic regression model. The identified predictors were used to create a clinically usable nomogram for identifying patients at risk of developing HUA. The area under curve (AUC) value of the final model was 0.692 for the training set and 0.706 for the test set. Important predictors of HUA were Zn, As, V, and Fe as well as consumption of wheat, beans, and rice; the corresponding estimated odds ratios and 95% confidence intervals were 1.091 (0.932,1.251), 1.190 (1.093,1.286), 0.924 (0.793,1.055), 0.704 (0.626,0.781), 0.998 (0.996,1.001), 0.993 (0.989,0.998), and 1.001 (0.998,1.002), respectively. In contrast to previous studies, we found that both urinary metal concentrations and dietary habits are important for predicting HUA risk. Exposure to specific metals and consumption of specific foods were identified as important predictors of HUA, indicating that the incidence of this disease could be reduced by reducing exposure to these metals and promoting improved dietary habits.

Functional impact of multi-omic interactions in breast cancer subtypes

Multi-omic approaches are expected to deliver a broader molecular view of cancer. However, the promised mechanistic explanations have not quite settled yet. Here, we propose a theoretical and computational analysis framework to semi-automatically produce network models of the regulatory constraints influencing a biological function. This way, we identified functions significantly enriched on the analyzed omics and described associated features, for each of the four breast cancer molecular subtypes. For instance, we identified functions sustaining over-representation of invasion-related processes in the basal subtype and DNA modification processes in the normal tissue. We found limited overlap on the omics-associated functions between subtypes; however, a startling feature intersection within subtype functions also emerged. The examples presented highlight new, potentially regulatory features, with sound biological reasons to expect a connection with the functions. Multi-omic regulatory networks thus constitute reliable models of the way omics are connected, demonstrating a capability for systematic generation of mechanistic hypothesis.

Genomic Effect of DNA Methylation on Gene Expression in Colorectal Cancer

Simple Summary

Abnormal DNA methylation is known to regulate gene expression, and its features have been frequently observed in colorectal cancer (CRC) patients. In addition, alterations in DNA methylation can be proposed as biomarkers for cancer prognosis, as they occur in the early stage of carcinogenesis. Although numerous studies have attempted to shed light on the impacts of DNA methylation on gene expression, it is still unclear which specific regions regulate gene expression and how they are associated with patient survival. In this study, we elucidated the intricate relationship between DNA methylation and gene expression. Furthermore, we found genes that were influenced by DNA methylation and were associated with survival; these genes were mainly enriched in immune-related pathways.

Abstract

The aberrant expression of cancer-related genes can lead to colorectal cancer (CRC) carcinogenesis, and DNA methylation is one of the causes of abnormal expression. Although many studies have been conducted to reveal how DNA methylation affects transcription regulation, the ways in which it modulates gene expression and the regions that significantly affect DNA methylation-mediated gene regulation remain unclear. In this study, we investigated how DNA methylation in specific genomic areas can influence gene expression. Several regression models were constructed for gene expression prediction based on DNA methylation. Among these models, ElasticNet, which had the best performance, was chosen for further analysis. DNA methylation near transcription start sites (TSS), especially from 2 kb upstream to 7 kb downstream of TSS, had an essential regulatory role in gene expression. Moreover, methylation-affected and survival-associated genes were compiled and found to be mainly enriched in immune-related pathways. This study investigated genomic regions in which methylation changes can affect gene expression. In addition, this study proposed that aberrantly expressed genes due to DNA methylation can lead to CRC pathogenesis by the immune system.

An epigenome-wide study of DNA methylation profiles and lung function among American Indians in the Strong Heart Study

BACKGROUND

Epigenetic modifications, including DNA methylation (DNAm), are often related to environmental exposures, and are increasingly recognized as key processes in the pathogenesis of chronic lung disease. American Indian communities have a high burden of lung disease compared to the national average. The objective of this study was to investigate the association of DNAm and lung function in the Strong Heart Study (SHS). We conducted a cross-sectional study of American Indian adults, 45-74 years of age who participated in the SHS. DNAm was measured using the Illumina Infinium Human MethylationEPIC platform at baseline (1989-1991). Lung function was measured via spirometry, including forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC), at visit 2 (1993-1995). Airflow limitation was defined as FEV1 < 70% predicted and FEV1/FVC < 0.7, restriction was defined as FEV1/FVC > 0.7 and FVC < 80% predicted, and normal spirometry was defined as FEV1/FVC > 0.7, FEV1 > 70% predicted, FVC > 80% predicted. We used elastic-net models to select relevant CpGs for lung function and spirometry-defined lung disease. We also conducted bioinformatic analyses to evaluate the biological plausibility of the findings.

RESULTS

Among 1677 participants, 21.2% had spirometry-defined airflow limitation and 13.6% had spirometry-defined restrictive pattern lung function. Elastic-net models selected 1118 Differentially Methylated Positions (DMPs) as predictors of airflow limitation and 1385 for restrictive pattern lung function. A total of 12 DMPs overlapped between airflow limitation and restrictive pattern. EGFR, MAPK1 and PRPF8 genes were the most connected nodes in the protein-protein interaction network. Many of the DMPs targeted genes with biological roles related to lung function such as protein kinases.

CONCLUSION

We found multiple differentially methylated CpG sites associated with chronic lung disease. These signals could contribute to better understand molecular mechanisms involved in lung disease, as assessed systemically, as well as to identify patterns that could be useful for diagnostic purposes. Further experimental and longitudinal studies are needed to assess whether DNA methylation has a causal role in lung disease.

Transcriptional, epigenetic and metabolic signatures in cardiometabolic syndrome defined by extreme phenotypes

BACKGROUND

This work is aimed at improving the understanding of cardiometabolic syndrome pathophysiology and its relationship with thrombosis by generating a multi-omic disease signature.

METHODS/RESULTS

We combined classic plasma biochemistry and plasma biomarkers with the transcriptional and epigenetic characterisation of cell types involved in thrombosis, obtained from two extreme phenotype groups (morbidly obese and lipodystrophy) and lean individuals to identify the molecular mechanisms at play, highlighting patterns of abnormal activation in innate immune phagocytic cells. Our analyses showed that extreme phenotype groups could be distinguished from lean individuals, and from each other, across all data layers. The characterisation of the same obese group, 6 months after bariatric surgery, revealed the loss of the abnormal activation of innate immune cells previously observed. However, rather than reverting to the gene expression landscape of lean individuals, this occurred via the establishment of novel gene expression landscapes. NETosis and its control mechanisms emerge amongst the pathways that show an improvement after surgical intervention.

CONCLUSIONS

We showed that the morbidly obese and lipodystrophy groups, despite some differences, shared a common cardiometabolic syndrome signature. We also showed that this could be used to discriminate, amongst the normal population, those individuals with a higher likelihood of presenting with the disease, even when not displaying the classic features.

Retrieval of Total Phosphorus Concentration in the Surface Water of Miyun Reservoir Based on Remote Sensing Data and Machine Learning Algorithms

Some essential water conservation areas in China have continuously suffered from various serious problems such as water pollution and water quality deterioration in recent decades and thus called for real-time water pollution monitoring system underwater resources management. On the basis of the remote sensing data and ground monitoring data, this study firstly constructed a more accurate retrieval model for total phosphorus (TP) concentration by comparing 12 machine learning algorithms, including support vector machine (SVM), artificial neural network (ANN), Bayesian ridge regression (BRR), lasso regression (Lasso), elastic net (EN), linear regression (LR), decision tree regressor (DTR), K neighbor regressor (KNR), random forest regressor (RFR), extra trees regressor (ETR), AdaBoost regressor (ABR) and gradient boosting regressor (GBR). Then, this study applied the constructed retrieval model to explore the spatial-temporal evolution of the Miyun Reservoir and finally assessed the water quality. The results showed that the model of TP concentration built by the ETR algorithm had the best accuracy, with the coefficient R2 reaching over 85% and the mean absolute error lower than 0.000433. The TP concentration in Miyun Reservoir was between 0.0380 and 0.1298 mg/L, and there was relatively significant spatial and temporal heterogeneity. It changed remarkably during the periods of the flood season, winter tillage, planting, and regreening, and it was lower in summer than in other seasons. Moreover, the TP in the southwest part of the reservoir was generally lower than in the northeast, as there was less human activities interference. According to the Environmental Quality Standard for the surface water environment, the water quality of Miyun Reservoir was overall safe, except only for an over-standard case occurrence in the spring and September. These conclusions can provide a significant scientific reference for water quality monitoring and management in Miyun Reservoir.

Multidimensional analysis of immune responses identified biomarkers of recent Mycobacterium tuberculosis infection

The risk of tuberculosis (TB) disease is higher in individuals with recent Mycobacterium tuberculosis (M.tb) infection compared to individuals with more remote, established infection. We aimed to define blood-based biomarkers to distinguish between recent and remote infection, which would allow targeting of recently infected individuals for preventive TB treatment. We hypothesized that integration of multiple immune measurements would outperform the diagnostic performance of a single biomarker. Analysis was performed on different components of the immune system, including adaptive and innate responses to mycobacteria, measured on recently and remotely M.tb infected adolescents. The datasets were standardized using variance stabilizing scaling and missing values were imputed using a multiple factor analysis-based approach. For data integration, we compared the performance of a Multiple Tuning Parameter Elastic Net (MTP-EN) to a standard EN model, which was built to the individual adaptive and innate datasets. Biomarkers with non-zero coefficients from the optimal single data EN models were then isolated to build logistic regression models. A decision tree and random forest model were used for statistical confirmation. We found no difference in the predictive performances of the optimal MTP-EN model and the EN model [average area under the receiver operating curve (AUROC) = 0.93]. EN models built to the integrated dataset and the adaptive dataset yielded identically high AUROC values (average AUROC = 0.91), while the innate data EN model performed poorly (average AUROC = 0.62). Results also indicated that integration of adaptive and innate biomarkers did not outperform the adaptive biomarkers alone (Likelihood Ratio Test χ2 = 6.09, p = 0.808). From a total of 193 variables, the level of HLA-DR on ESAT6/CFP10-specific Th1 cytokine-expressing CD4 cells was the strongest biomarker for recent M.tb infection. The discriminatory ability of this variable was confirmed in both tree-based models. A single biomarker measuring M.tb-specific T cell activation yielded excellent diagnostic potential to distinguish between recent and remote M.tb infection.

Incorporating prior knowledge into regularized regression

MOTIVATION

Associated with genomic features like gene expression, methylation and genotypes, used in statistical modeling of health outcomes, there is a rich set of meta-features like functional annotations, pathway information and knowledge from previous studies, that can be used post hoc to facilitate the interpretation of a model. However, using this meta-feature information a priori rather than post hoc can yield improved prediction performance as well as enhanced model interpretation.

RESULTS

We propose a new penalized regression approach that allows a priori integration of external meta-features. The method extends LASSO regression by incorporating individualized penalty parameters for each regression coefficient. The penalty parameters are, in turn, modeled as a log-linear function of the meta-features and are estimated from the data using an approximate empirical Bayes approach. Optimization of the marginal likelihood on which the empirical Bayes estimation is performed using a fast and stable majorization-minimization procedure. Through simulations, we show that the proposed regression with individualized penalties can outperform the standard LASSO in terms of both parameters estimation and prediction performance when the external data is informative. We further demonstrate our approach with applications to gene expression studies of bone density and breast cancer.

AVAILABILITY AND IMPLEMENTATION

The methods have been implemented in the R package xtune freely available for download from https://cran.r-project.org/web/packages/xtune/index.html.

Development of a novel risk score for the prediction of critical illness amongst COVID-19 patients

OBJECTIVES

Coronavirus disease-19 (COVID-19) is associated with various clinical manifestations, ranging from asymptomatic infection to critical illness. The aim of this study is to evaluate the clinical and laboratory characteristics of hospitalised COVID-19 patients and construct a predictive model for the discrimination of patients at risk of disease progression.

METHODS

A single-centre cohort study was conducted including consecutively patients with COVID-19. Demographic, clinical and laboratory findings were prospectively collected at admission. The primary outcome of interest was the intensive care unit admission. A risk model was constructed by applying a Cox's proportional hazard's model with elastic net penalty. Its diagnostic performance was assessed by receiver operating characteristic analysis and was compared with conventional pneumonia severity scores.

RESULTS

From a total of 67 patients 15 progressed to critical illness. The risk score included patients' gender, presence of hypertension and diabetes mellitus, fever, shortness of breath, serum glucose, aspartate aminotransferase, lactate dehydrogenase, C-reactive protein and fibrinogen. Its predictive accuracy was estimated to be high (area under the curve: 97.1%), performing better than CURB-65, CRB-65 and PSI/PORT scores. Its sensitivity and specificity were estimated to be 92.3% and 93.3%, respectively, at the optimal threshold of 1.6.

CONCLUSIONS

A10-variable risk score was constructed based on clinical and laboratory characteristics in order to predict critical illness amongst hospitalised COVID-19 patients, achieving better discrimination compared with traditional pneumonia severity scores. The proposed risk model should be externally validated in independent cohorts in order to ensure its prognostic efficacy.

Multi-Omic Regulation of the PAM50 Gene Signature in Breast Cancer Molecular Subtypes

Breast cancer is a disease that exhibits heterogeneity that goes from the genomic to the clinical levels. This heterogeneity is thought to be captured (at least partially) by the so-called breast cancer molecular subtypes. These molecular subtypes were initially defined based on the unsupervised clustering of gene expression and its correlate with histological, morphological, phenotypic and clinical features already known. Later, a 50-gene signature, PAM50, was defined in order to identify the biological subtype of a given sample within the clinical setting. The PAM50 signature was obtained by the use of unsupervised statistical methods, and therefore no limitation was set on the biological relevance (or lack of) of the selected genes beyond its predictive capacity. An open question that remains is what are the regulatory elements that drive the various expression behaviors of this set of genes in the different molecular subtypes. This question becomes more relevant as the measurement of more biological layers of regulation becomes accessible. In this work, we analyzed the gene expression regulation of the 50 genes in the PAM50 signature, in terms of (a) gene co-expression, (b) transcription factors, (c) micro-RNAs, and (d) methylation. Using data from the Cancer Genome Atlas (TCGA) for the Luminal A and B, Basal, and HER2-enriched molecular subtypes as well as normal tumor adjacent tissue, we identified predictors for gene expression through the use of an elastic net model. We compare and contrast the sets of identified regulators for the gene signature in each molecular subtype, and systematically compare them to current literature. We also identified a unique set of predictors for the expression of genes in the PAM50 signature associated with each of the molecular subtypes. Most selected predictors are exclusive for a PAM50 gene and predictors are not shared across subtypes. There are only 13 coding transcripts and 2 miRNAs selected for the four subtypes. MiR-21 and miR-10b connect almost all the PAM50 genes in all the subtypes and normal tissue, but do it in an exclusive manner, suggesting a cancer switch from miR-10b coordination in normal tissue to miR-21. The PAM50 gene sets of selected predictors that enrich for a function across subtypes, support that different regulatory molecular mechanisms are taking place. With this study we aim to a wider understanding of the regulatory mechanisms that differentiate the expression of the PAM50 signature, which in turn could perhaps help understand the molecular basis of the differences between the molecular subtypes.

The Many Faces of Gene Regulation in Cancer: A Computational Oncogenomics Outlook

Cancer is a complex disease at many different levels. The molecular phenomenology of cancer is also quite rich. The mutational and genomic origins of cancer and their downstream effects on processes such as the reprogramming of the gene regulatory control and the molecular pathways depending on such control have been recognized as central to the characterization of the disease. More important though is the understanding of their causes, prognosis, and therapeutics. There is a multitude of factors associated with anomalous control of gene expression in cancer. Many of these factors are now amenable to be studied comprehensively by means of experiments based on diverse omic technologies. However, characterizing each dimension of the phenomenon individually has proven to fall short in presenting a clear picture of expression regulation as a whole. In this review article, we discuss some of the more relevant factors affecting gene expression control both, under normal conditions and in tumor settings. We describe the different omic approaches that we can use as well as the computational genomic analysis needed to track down these factors. Then we present theoretical and computational frameworks developed to integrate the amount of diverse information provided by such single-omic analyses. We contextualize this within a systems biology-based multi-omic regulation setting, aimed at better understanding the complex interplay of gene expression deregulation in cancer.