Multiplex-Heterogeneous Network-Based Capturing Potential SNP "Switches" of Pathways Associating With Diverse Disease Characteristics of Asthma

Abstract

Asthma is a complex heterogeneous respiratory disorder. In recent years nubbly regions of the role of genetic variants and transcriptome including mRNAs, microRNAs, and long non-coding RNAs in the pathogenesis of asthma have been separately excavated and reported. However, how to systematically integrate and decode this scattered information remains unclear. Further exploration would improve understanding of the internal communication of asthma. To excavate new insights into the pathogenesis of asthma, we ascertained three asthma characteristics according to reviews, airway inflammation, airway hyperresponsiveness, and airway remodeling. We manually created a contemporary catalog of corresponding risk transcriptome, including mRNAs, miRNAs, and lncRNAs. MIMP is a multiplex-heterogeneous networks-based approach, measuring the relevance of disease characteristics to the pathway by examining the similarity between the determined vectors of risk transcriptome and pathways in the same low-dimensional vector space. It was developed to enable a more concentrated and in-depth exploration of potential pathways. We integrated experimentally validated competing endogenous RNA regulatory information and the SNPs with significant pathways into the ceRNA-mediated SNP switching pathway network (CSSPN) to analyze ceRNA regulation of pathways and the role of SNP in these dysfunctions. We discovered 11 crucial ceRNA regulations concerning asthma disease feature pathway and propose a potential mechanism of ceRNA regulatory SNP → gene → pathway → disease feature effecting asthma pathogenesis, especially for MALAT1 (rs765499057/rs764699354/rs189435941) → hsa-miR-155 → IL13 (rs201185816/rs1000978586/rs202101165) → Interleukin-4 and Interleukin-13 signaling → inflammation/airway remodeling and MALAT1 (rs765499057/rs764699354/rs189435941) → hsa-miR-155 → IL17RB (rs948046241) → Interleukin-17 signaling (airway remodeling)/Cytokine-cytokine receptor interaction (inflammation). This study showed a systematic and propagable workflow for capturing the potential SNP "switch" of asthma through text and database mining and provides further information on the pathogenesis of asthma.

Keywords: SNPs; asthma; bioinformatics; ceRNA; lncRNA; miRNA; network; pathway enrichment analysis.

FIGURE 1

Compiling the disease feature risk transcriptome. (A) The distribution of disease characteristics-associated genes. (B) Spatial distribution of disease characteristics-related genes on chromosomes. (C) The top five statistically significant Go term annotations of disease characteristics-associated genes and the similarity measuring.

FIGURE 2

The structure of MIMP and pathway result exhibition. (A) The structure illustration of complex multiplex-heterogeneous network embedding in MIMP. (B) Robust risk pathways of disease characteristics. The number of genes mapping to pathways was directly labeled and the distribution of diverse gene types was color-coded. Pathways that were statistically significant in multiple disease features are highlighted via colored names and linked to the pie plot. The vertical axis from bottom to top shows the ordering of the empirical p-values from large to small, that statistical significance tends to increase.

FIGURE 3

The comprehensive disease regulatory network and information decoding. (A) CSSPN. Green circle, orange rhombuses, and blue inverted triangles represent mRNAs, miRNAs, and lncRNAs respectively. The SNPs located in lncRNA, pri-/pre-miRNA, and the 3′UTR of mRNA were shown with the red circle around the mRNAs, miRNAs, and lncRNAs. The SNPs within the 3′UTR of mRNA, the seed region of mature miRNA, and the miRNA binding site on lncRNA were exhibited with green, orange, and blue dash lines respectively. The red dash line indicates the seed region of mature miRNA and 3′UTR of mRNA contain SNPs, which could affect miRNA-mRNA binding. (B) The multi-bar plot shows gene mapping and SNP mapping information pathways. The top and bottom axis represented the number of mapping genes and mapping SNPs respectively. (C) The dot plot of regulatory SNPs density and ceRNA regulation with SNP ratio. The top axis showed the density of regulatory SNPs. The percentage represented by the bottom axis reflects the proportion of ceRNA regulation with SNP in the total ceRNA regulations.

FIGURE 4

A depiction of the disease features of the ceRNA network-involved pathway. Proteins encoded by risk genes concerning inflammation, hyperresponsiveness, and remodeling and related complexes are indicated in orange, green, and blue respectively, while those in red character represented encoding genes indicate the ceRNA regulation-containing genes with regulatory SNPs. Yellow represents the significant elements related to all disease features in a pathway.

FIGURE 5

The schematic diagram of regulatory SNPs → gene → pathway effect of focused genes. The green and blue represented the genes in the risk pathways concerning inflammation and remodeling respectively. The diamond, circle, and rectangle represent lncRNA, miRNA, and mRNA respectively. The line between them shows the regulation of ceRNA. The “on-off” symbol sited on line between lncRNA and miRNA or miRNA and mRNA indicate the regulatory SNP within the miRNA binding site on lncRNA or within the mature miRNA seed region (character colored in deep grey) or the 3′UTR region (character colored in black). The peripheral large circles or rectangles denoted the pathways that regulatory SNPs may influence through their effects on ceRNA regulations to target genes.

FIGURE 6

The potential mechanisms of regulatory SNPs influence IL13 and IL17RB via ceRNA regulation.