Identifying Methylation Signatures and Rules for COVID-19 With Machine Learning Methods

Exploring TSPAN4 promoter methylation as a diagnostic biomarker for tuberculosis

Background

Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is a persistent infectious disease threatening human health. The existing diagnostic methods still have significant shortcomings, including a low positivity rate in pathogen-based diagnoses and the inability of immunological diagnostics to detect active TB. Hence, it is urgent to develop new techniques to detect TB more accurate and earlier. This research aims to scrutinize and authenticate DNA methylation markers suitable for tuberculosis diagnosis. Concurrently, Providing a new approach for tuberculosis diagnosis.

Methods

Blood samples from patients with newly diagnosed tuberculosis and healthy controls (HC) were utilized in this study. Examining methylation microarray data from 40 whole blood samples (22TB + 18HC), we employed two procedures: signature gene methylated position analysis and signature region methylated position analysis to pinpoint distinctive methylated positions. Based on the screening results, diagnostic classifiers are constructed through machine learning, and validation was conducted through pyrosequencing in a separate queue (22TB + 18HC). Culminating in the development of a new tuberculosis diagnostic method via quantitative real-time methylation specific PCR (qMSP).

Results

The combination of the two procedures revealed a total of 10 methylated positions, all of which were located in the promoter region. These 10 signature methylated positions facilitated the construction of a diagnostic classifier, exhibiting robust diagnostic accuracy in both cross-validation and external test sets. The LDA model demonstrated the best classification performance, achieving an AUC of 0.83, specificity of 0.8, and sensitivity of 0.86 on the external test set. Furthermore, the validation of signature methylated positions through pyrosequencing demonstrated high agreement with screening outcomes. Additionally, qMSP detection of 2 potential hypomethylated positions (cg04552852 and cg12464638) exhibited promising results, yielding an AUC of 0.794, specificity of 0.720, and sensitivity of 0.816.

Conclusion

Our study demonstrates that the validated signature methylated positions through pyrosequencing emerge as plausible biomarkers for tuberculosis diagnosis. The specific methylation markers in the TSPAN4 gene, identified in whole blood samples, hold promise for improving tuberculosis diagnosis. This approach could significantly enhance diagnostic accuracy and speed, offering a new avenue for early detection and treatment.

Identification of biomarkers and pathways for the SARS-CoV-2 infections in obstructive sleep apnea patients based on machine learning and proteomic analysis

BACKGROUND

The prevalence of obstructive sleep apnea (OSA) was found to be higher in individuals following COVID-19 infection. However, the intricate mechanisms that underscore this concomitance remain partially elucidated. The aim of this study was to delve deeper into the molecular mechanisms that underpin this comorbidity.

METHODS

We acquired gene expression profiles for COVID-19 (GSE157103) and OSA (GSE75097) from the Gene Expression Omnibus (GEO) database. Upon identifying shared feature genes between OSA and COVID-19 utilizing LASSO, Random forest and Support vector machines algorithms, we advanced to functional annotation, analysis of protein-protein interaction networks, module construction, and identification of pivotal genes. Furthermore, we established regulatory networks encompassing transcription factor (TF)-gene and TF-miRNA interactions, and searched for promising drug targets. Subsequently, the expression levels of pivotal genes were validated through proteomics data from COVID-19 cases.

RESULTS

Fourteen feature genes shared between OSA and COVID-19 were selected for further investigation. Through functional annotation, it was indicated that metabolic pathways play a role in the pathogenesis of both disorders. Subsequently, employing the cytoHubba plugin, ten hub genes were recognized, namely TP53, CCND1, MDM2, RB1, HIF1A, EP300, STAT3, CDK2, HSP90AA1, and PPARG. The finding of proteomics unveiled a substantial augmentation in the expression level of HSP90AA1 in COVID-19 patient samples, especially in severe conditions.

CONCLUSIONS

Our investigation illuminate a mutual pathogenic mechanism that underlies both OSA and COVID-19, which may provide novel perspectives for future investigations into the underlying mechanisms.

Systems biology approaches to identify driver genes and drug combinations for treating COVID-19

Corona virus 19 (Covid-19) has caused many problems in public health, economic, and even cultural and social fields since the beginning of the epidemic. However, in order to provide therapeutic solutions, many researches have been conducted and various omics data have been published. But there is still no early diagnosis method and comprehensive treatment solution. In this manuscript, by collecting important genes related to COVID-19 and using centrality and controllability analysis in PPI networks and signaling pathways related to the disease; hub and driver genes have been identified in the formation and progression of the disease. Next, by analyzing the expression data, the obtained genes have been evaluated. The results show that in addition to the significant difference in the expression of most of these genes, their expression correlation pattern is also different in the two groups of COVID-19 and control. Finally, based on the drug-gene interaction, drugs affecting the identified genes are presented in the form of a bipartite graph, which can be used as the potential drug combinations.

Immune responses of different COVID-19 vaccination strategies by analyzing single-cell RNA sequencing data from multiple tissues using machine learning methods

Multiple types of COVID-19 vaccines have been shown to be highly effective in preventing SARS-CoV-2 infection and in reducing post-infection symptoms. Almost all of these vaccines induce systemic immune responses, but differences in immune responses induced by different vaccination regimens are evident. This study aimed to reveal the differences in immune gene expression levels of different target cells under different vaccine strategies after SARS-CoV-2 infection in hamsters. A machine learning based process was designed to analyze single-cell transcriptomic data of different cell types from the blood, lung, and nasal mucosa of hamsters infected with SARS-CoV-2, including B and T cells from the blood and nasal cavity, macrophages from the lung and nasal cavity, alveolar epithelial and lung endothelial cells. The cohort was divided into five groups: non-vaccinated (control), 2*adenovirus (two doses of adenovirus vaccine), 2*attenuated (two doses of attenuated virus vaccine), 2*mRNA (two doses of mRNA vaccine), and mRNA/attenuated (primed by mRNA vaccine, boosted by attenuated vaccine). All genes were ranked using five signature ranking methods (LASSO, LightGBM, Monte Carlo feature selection, mRMR, and permutation feature importance). Some key genes that contributed to the analysis of immune changes, such as RPS23, DDX5, PFN1 in immune cells, and IRF9 and MX1 in tissue cells, were screened. Afterward, the five feature sorting lists were fed into the feature incremental selection framework, which contained two classification algorithms (decision tree [DT] and random forest [RF]), to construct optimal classifiers and generate quantitative rules. Results showed that random forest classifiers could provide relative higher performance than decision tree classifiers, whereas the DT classifiers provided quantitative rules that indicated special gene expression levels under different vaccine strategies. These findings may help us to develop better protective vaccination programs and new vaccines.

Identification of dynamic gene expression profiles during sequential vaccination with ChAdOx1/BNT162b2 using machine learning methods

To date, COVID-19 remains a serious global public health problem. Vaccination against SARS-CoV-2 has been adopted by many countries as an effective coping strategy. The strength of the body’s immune response in the face of viral infection correlates with the number of vaccinations and the duration of vaccination. In this study, we aimed to identify specific genes that may trigger and control the immune response to COVID-19 under different vaccination scenarios. A machine learning-based approach was designed to analyze the blood transcriptomes of 161 individuals who were classified into six groups according to the dose and timing of inoculations, including I-D0, I-D2-4, I-D7 (day 0, days 2–4, and day 7 after the first dose of ChAdOx1, respectively) and II-D0, II-D1-4, II-D7-10 (day 0, days 1–4, and days 7–10 after the second dose of BNT162b2, respectively). Each sample was represented by the expression levels of 26,364 genes. The first dose was ChAdOx1, whereas the second dose was mainly BNT162b2 (Only four individuals received a second dose of ChAdOx1). The groups were deemed as labels and genes were considered as features. Several machine learning algorithms were employed to analyze such classification problem. In detail, five feature ranking algorithms (Lasso, LightGBM, MCFS, mRMR, and PFI) were first applied to evaluate the importance of each gene feature, resulting in five feature lists. Then, the lists were put into incremental feature selection method with four classification algorithms to extract essential genes, classification rules and build optimal classifiers. The essential genes, namely, NRF2, RPRD1B, NEU3, SMC5, and TPX2, have been previously associated with immune response. This study also summarized expression rules that describe different vaccination scenarios to help determine the molecular mechanism of vaccine-induced antiviral immunity.