An Application of Sequential Meta-Analysis to Gene Expression Studies

The Role of Systems Biology in Deciphering Asthma Heterogeneity

Asthma is one of the most common and lifelong and chronic inflammatory diseases characterized by inflammation, bronchial hyperresponsiveness, and airway obstruction episodes. It is a heterogeneous disease of varying and overlapping phenotypes with many confounding factors playing a role in disease susceptibility and management. Such multifactorial disorders will benefit from using systems biology as a strategy to elucidate molecular insights from complex, quantitative, massive clinical, and biological data that will help to understand the underlying disease mechanism, early detection, and treatment planning. Systems biology is an approach that uses the comprehensive understanding of living systems through bioinformatics, mathematical, and computational techniques to model diverse high-throughput molecular, cellular, and the physiologic profiling of healthy and diseased populations to define biological processes. The use of systems biology has helped understand and enrich our knowledge of asthma heterogeneity and molecular basis; however, such methods have their limitations. The translational benefits of these studies are few, and it is recommended to reanalyze the different studies and omics in conjugation with one another which may help understand the reasons for this variation and help overcome the limitations of understanding the heterogeneity in asthma pathology. In this review, we aim to show the different factors that play a role in asthma heterogeneity and how systems biology may aid in understanding and deciphering the molecular basis of asthma.

Derangement of cell cycle markers in peripheral blood mononuclear cells of asthmatic patients as a reliable biomarker for asthma control

In asthma, most of the identified biomarkers pertain to the Th2 phenotype and no known biomarkers have been verified for severe asthmatics. Therefore, identifying biomarkers using the integrative phenotype-genotype approach in severe asthma is needed. The study aims to identify novel biomarkers as genes or pathways representing the core drivers in asthma development, progression to the severe form, resistance to therapy, and tissue remodeling regardless of the sample cells or tissues examined. Comprehensive reanalysis of publicly available transcriptomic data that later was validated in vitro, and locally recruited patients were used to decipher the molecular basis of asthma. Our in-silicoanalysis revealed a total of 10 genes (GPRC5A, SFN, ABCA1, KRT8, TOP2A, SERPINE1, ANLN, MKI67, NEK2, and RRM2) related to cell cycle and proliferation to be deranged in the severe asthmatic bronchial epithelium and fibroblasts compared to their healthy counterparts. In vitro, RT qPCR results showed that (SERPINE1 and RRM2) were upregulated in severe asthmatic bronchial epithelium and fibroblasts, (SFN, ABCA1, TOP2A, SERPINE1, MKI67, and NEK2) were upregulated in asthmatic bronchial epithelium while (GPRC5A and KRT8) were upregulated only in asthmatic bronchial fibroblasts. Furthermore, MKI76, RRM2, and TOP2A were upregulated in Th2 high epithelium while GPRC5A, SFN, ABCA1 were upregulated in the blood of asthmatic patients. SFN, ABCA1 were higher, while MKI67 was lower in severe asthmatic with wheeze compared to nonasthmatics with wheezes. SERPINE1 and GPRC5A were downregulated in the blood of eosinophilic asthmatics, while RRM2 was upregulated in an acute attack of asthma. Validation of the gene expression in PBMC of locally recruited asthma patients showed that SERPINE1, GPRC5A, SFN, ABCA1, MKI67, and RRM2 were downregulated in severe uncontrolled asthma. We have identified a set of biologically crucial genes to the homeostasis of the lung and in asthma development and progression. This study can help us further understand the complex interplay between the transcriptomic data and the external factors which may deviate our understanding of asthma heterogeneity.

Meta-analysis approach as a gene selection method in class prediction: does it improve model performance? A case study in acute myeloid leukemia

BACKGROUND

Aggregating gene expression data across experiments via meta-analysis is expected to increase the precision of the effect estimates and to increase the statistical power to detect a certain fold change. This study evaluates the potential benefit of using a meta-analysis approach as a gene selection method prior to predictive modeling in gene expression data.

RESULTS

Six raw datasets from different gene expression experiments in acute myeloid leukemia (AML) and 11 different classification methods were used to build classification models to classify samples as either AML or healthy control. First, the classification models were trained on gene expression data from single experiments using conventional supervised variable selection and externally validated with the other five gene expression datasets (referred to as the individual-classification approach). Next, gene selection was performed through meta-analysis on four datasets, and predictive models were trained with the selected genes on the fifth dataset and validated on the sixth dataset. For some datasets, gene selection through meta-analysis helped classification models to achieve higher performance as compared to predictive modeling based on a single dataset; but for others, there was no major improvement. Synthetic datasets were generated from nine simulation scenarios. The effect of sample size, fold change and pairwise correlation between differentially expressed (DE) genes on the difference between MA- and individual-classification model was evaluated. The fold change and pairwise correlation significantly contributed to the difference in performance between the two methods. The gene selection via meta-analysis approach was more effective when it was conducted using a set of data with low fold change and high pairwise correlation on the DE genes.

CONCLUSION

Gene selection through meta-analysis on previously published studies potentially improves the performance of a predictive model on a given gene expression data.