Identification and interaction analysis of molecular markers in myocardial infarction by bioinformatics and next-generation sequencing data analysis

Abstract

Background

Cardiovascular diseases are prevalent worldwide with any age, and it is characterized by sudden blockage of blood flow to heart and permanent damage to the heart muscle, whose cause and underlying molecular mechanisms are not fully understood. This investigation aimed to explore and identify essential genes and signaling pathways that contribute to the progression of MI.

Methods

The aim of this investigation was to use bioinformatics and next-generation sequencing (NGS) data analysis to identify differentially expressed genes (DEGs) with diagnostic and therapeutic potential in MI. NGS dataset (GSE132143) was downloaded from the Gene Expression Omnibus (GEO) database. DEGs between MI and normal control samples were identified using the DESeq2 R bioconductor tool. The gene ontology (GO) and REACTOME pathway enrichment analyses of the DEGs were performed using g:Profiler. Next, four kinds of algorithms in the protein–protein interaction (PPI) were performed to identify potential novel biomarkers. Next, miRNA-hub gene regulatory network analysis and TF-hub gene regulatory network were constructed by miRNet and NetworkAnalyst database, and Cytoscape software. Finally, the diagnostic effectiveness of hub genes was predicted by receiver operator characteristic curve (ROC) analysis and AUC more than 0.800 was considered as having the capability to diagnose MI with excellent specificity and sensitivity.

Results

A total of 958 DEGs were identified, consisting of 480 up-regulated genes and 478 down-regulated genes. The enriched GO terms and pathways of the DEGs include immune system, neuronal system, response to stimulus and multicellular organismal process. Ten hub genes (namely cftr, cdk1, rps13, rps15a, rps27, notch1, mrpl12, nos2, ccdc85b and atn1) were obtained via protein–protein interaction analysis results. MiRNA-hub gene regulatory network and TF-hub gene regulatory network showed that hsa-mir-409-3p, hsa-mir-3200-3p, creb1 and tp63 might play an important role in the MI.

Conclusions

Analysis of next-generation sequencing dataset combined with global network information and validation presents a successful approach to uncover the risk hub genes and prognostic markers of MI. Our investigation identified four risk- and prognostic-related gene signatures, including cftr, cdk1, rps13, rps15a, rps27, notch1, mrpl12, nos2, ccdc85b and atn1. This gene sets contribute a new perspective to improve the diagnostic, prognostic, and therapeutic outcomes of MI.

Background

Myocardial infarction (MI) is one of the most commonly diagnosed cardiovascular diseases prevalent worldwide [1]. The clinical prevalence of MI is approximately 18 million people worldwide [1], and its main features such as the presence of sudden blockage of blood flow to heart and permanent damage to the heart muscle [2]. MI becomes the chief cause of morbidity and mortality all over the world [3]. Studies have revealed that the progression of MI is related to genetic factors [4], older age [5], smoking [6], hypertension [7], diabetes mellitus [8] and obesity [9]. However, the molecular pathogenesis of MI has not been explored comprehensively. In current years, molecular biomarkers were demonstrated highly useful as clinical tools for MI diagnosis [10, 11]. A key molecular target for diagnosis and reexamination is essential for therapeutic action and prognostic outcome of MI patients.

The diagnosis of MI, however, remains a challenge due to the complex clinical features and the lack of definitive diagnostic tests or biomarkers in the early stages. Several studies have described that significant molecular biomarkers in MI were identified as prognostic or therapeutic factors such as hmga1 [12], stxbp2 [13], ccl19 and ccl21 [14], nlrp3 [15] and cd14 [16]. Molecular biology investigation has identified numerous signaling pathways that contribute to the progression of MI, including, Nrf2/HO-1; TLR4/TNF-α signaling pathway [17], Wnt/β-catenin signaling pathway [18], β2AR, cAMP/PKA, and BDNF/TrkB signaling pathways [19], PTEN/PI3K/Akt signaling pathway [20] and TGF-β1/SMAD2 signaling pathway [21]. It is very essential to explore the molecular characteristics and mechanism of MI occurrence, advancement to contribute novel approach for the effective prevention, diagnosis and treatment of MI.

In current investigation, next-generation sequencing (NGS) technology based on high-throughput platforms has been extensively used to scrutinize and determine the encouraging biomarkers for diagnosis and prognosis of disease at the genome level. With the wide application of bioinformatics and NGS data analysis, a huge number of cores data have been produced, and most of the NGS data have been deposited and stored in public database Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) [24]. Bioinformatics and NGS data analysis have been carried out on cardiovascular diseases in recently years [22], and hundreds of differentially expressed genes (DEGs) have been obtained. In perspective of this, the plan of this investigation is to analyze NGS data related to the pathogenesis of MI from the public database, which provide novel understanding into the molecular mechanisms of MI.

Here, we analyzed the MI NGS dataset GSE132143 [23] was downloaded from the GEO database and performed the identification of differentially expressed genes (DEGs) between the MI samples and normal control samples. Afterward, gene ontology (GO) terms and REACTOME pathways linked with DEGs were marked to illuminate the gene enrichment in MI. A protein–protein interaction network (PPIN) was generated to resolve the hub genes linked with MI. Moreover, miRNA-hub gene regulatory network analysis and TF-hub gene regulatory network, and the miRNA and TFs were identified according to degree parameters. Finally, hub genes were validated via receiver operating characteristic curve (ROC) analysis. In brief, these outcomes determined that the most genes might have probable value in prognosis and diagnosis of MI.

Methods

Data resources

The GEO database is a national depository of genetic information databases of next-generation sequencing data. In this investigation, we retrieved NGS dataset GSE132143 [23] from the GEO database. GSE132143 (Platform GPL18573 Illumina NextSeq 500 (Homo sapiens)) comprised 20 MI samples and 12 normal control samples.

Identification of DEGs

DESeq2 of R/Bioconductor software package [25] was accessible to determine the DEGs between MI and normal control samples. The log-fold change (FC) and adjusted P-values (adj. P) were resolved. The adj. P using the Benjamini–Hochberg method with default values was enforced to accurate the probable false-positive results [26]. According standard statistical protocol or literature (logFC > 1 and logFc < −1) and adj.p.Val < 0.05 are significant up- and down-regulated genes selection. Genes that met the definite cut-off criteria of adj. P < 0.05 and |logFC|> 1.37 were marked as up-regulated DEGs and |logFC|< −1.89 was marked as down-regulated DEGs. We used a ggplot2 and gplot package of R/Bioconductor software to generate volcano plots and heatmap of the DEGs.

GO and REACTOME pathway enrichment analysis of DEGs

The DEGs with the strongest associations were selected for GO and REACTOME pathway enrichment analysis. Moreover, GO and REACTOME pathway enrichment analysis were used to further determine the potential functions of the essential genes associated in MI. To effectively examine the biological content behind DEGs, the g:Profiler (http://biit.cs.ut.ee/gprofiler/) [29] was used to gain the set of functional annotation. GO enrichment analysis (GO, http://www.geneontology.org) [27] is a broadly used tool for annotating genes with probable functions, such as biological process (BP), cellular component (CC), and molecular function (MF). REACTOME (https://reactome.org/) [28] pathway enrichment analysis is an efficient resource for analytical investigation of gene functions and linked high-level genome functional information. p < 0.05 was marked as the threshold value.

Construction of the PPI network and module analysis

The Human Integrated Protein–Protein Interaction rEference (HIPPIE) (http://cbdm-01.zdv.uni-mainz.de/~mschaefer/hippie/) [30] database is an online search tool used to analyze accepted proteins and predict PPI networks, including direct and indirect interactions between proteins and their functional correlations and then software Cytoscape (version 3.8.2) (http://www.cytoscape.org/) [31] was engaged to accommodate and visualize PPI networks. HIPPIE is an experimentally validated PPI database. PPI networks can advance the examination of molecular targets, signaling pathways, and network functions associated in the advancement of MI. Hub genes are associated biological processes and diseases progression. Subsequently, we utilized a plug-in Network Analyzer of Cytoscape to resolve hub genes according to the algorithms, including node degree [32], betweenness centrality [33], stress centrality [34] and closeness centrality [35]. The PEWCC1 [36] plug-in was used to cover the hub gene modules in the PPI network.

MiRNA-hub gene regulatory network construction

MicroRNAs can inflect hub gene expression by promoting or inhibiting mRNA degradation and translation. Target hub gene information of miRNAs was collected from miRNet database (https://www.mirnet.ca/) [37], which is an experimentally validated miRNA-hub gene interactions database. The intersection of target hub genes of miRNAs in MI was used to construct the miRNA-hub gene-regulated network. The miRNA-hub gene regulatory network of the hub genes and their targeted miRNAs was visualized by Cytoscape software [31].

TF-hub gene regulatory network construction

Transcription factors (TFs) can inflect hub gene expression by promoting mRNA translation. Target hub gene information of TFs was collected from NetworkAnalyst database (https://www.networkanalyst.ca/) [38], which is an experimentally validated TF-hub gene interactions database. The intersection of target hub genes of TFs in MI was used to construct the TF-hub gene-regulated network. The TF-hub gene regulatory network of the hub genes and their targeted TFs was visualized by Cytoscape software [31].

Validation of hub genes by receiver operating characteristic curve (ROC) analysis

The linear discriminant analysis (LDA) with combined selected hub genes was used to identify biomarkers with high sensitivity and specificity for MI diagnosis. Here, we used one data as training dataset and other data as validation dataset iteratively. The receiver operator characteristic curves were plotted, and area under curve (AUC) was determined separately to check the achievement of LDA model using the R packages “pROC” [39]. A AUC > 0.8 determined that the model had a favorable fitting effect.

Results

Identification of DEGs

In our investigation, 958 DEGs were identified between MI samples and normal control samples. Among them, 480 were up-regulated genes (|logFC|> 1.37) and 478 were down-regulated genes (|logFC|< −1.89) (Table 1). The volcano plot (Fig. 1) was used to show the expression pattern of DEGs in MI (volcano plots presented the distributions of differentially expressed genes; Green dots—up-regulated genes and red dots—down-regulated genes). The heatmap of the DEGs (heat maps explains the gene expression levels in dataset) is shown in Fig. 2.

Table 1 The statistical metrics for key differentially expressed genes (DEGs)

Fig. 1

Volcano plot of differentially expressed genes. Genes with a significant change of more than twofold were selected. Green dot represents up-regulated significant genes, and red dot represents down-regulated significant genes

Fig. 2

Heat map of differentially expressed genes. Legend on the top left indicates log-fold change of genes. (A1–A20 = MI samples; B1–B12 = normal control samples)

GO and REACTOME pathway enrichment analysis of DEGs

A total of 480 up-regulated genes and 478 down-regulated genes were analyzed by g:Profiler software. GO enrichment analysis covers three aspects: BP, CC and MF (Table 2). The up-regulated genes were mainly related to response to stimulus, cell communication, cell periphery, membrane, molecular transducer activity and transmembrane signaling receptor activity; while the down-regulated genes were mainly involved in multicellular organismal process, anatomical structure development, cell periphery, plasma membrane, inorganic molecular entity transmembrane transporter activity and structural molecule activity. Moreover, REACTOME pathway enrichment analysis indicated that the up-regulated genes were involved in immune system and innate immune system, while down-regulated genes were involved in neuronal system and extracellular matrix organization (Table 3).

Table 2 The enriched GO terms of the up- and down-regulated differentially expressed genes

Table 3 The enriched pathway terms of the up- and down-regulated differentially expressed genes

Construction of the PPI network and module analysis

Considering the critical role of protein interactions in protein function, we used the HIPIE database and Cytoscape software to generate PPI network once we had identified the 958 DEGs. The results showed that there were dense regions in PPI, that is, genes closely related to MI. A total of 6292 nodes and 12,582 edges were selected to plot the PPI network (Fig. 3). The Network Analyzer plugin of Cytoscape was used to score each nodes by 4 selected algorithms, including node degree, betweenness, stress and closeness. Finally, we identified ten hub genes (cftr, cdk1, rps13, rps15a, rps27, notch1, mrpl12, nos2, ccdc85b and atn1) and are listed Table 4. Then, the significant modules were identified via the PEWCC1 plugin. The top two significant modules were selected. The functional enrichment analysis of genes in module 1 and module 2 were conducted by g:Profiler. Module 1 consisted of 59 nodes and 203 edges (Fig. 4A). Hub genes in module 1 were significantly enriched in metabolism of amino acids and derivatives, cellular responses to external stimuli, and membrane. Module 2 consisted of 21 nodes and 41 edges (Fig. 4B). Hub genes in module 2 were significantly enriched in diseases of glycosylation and multicellular organismal process.

Fig. 3

PPI network of DEGs. Up-regulated genes are marked in green; down-regulated genes are marked in red

Table 4 Topology table for up- and down-regulated genes

Fig. 4

Modules of isolated form PPI of DEGs. (A) The most significant module was obtained from PPI network with 59 nodes and 203 edges for up-regulated genes (B). The most significant module was obtained from PPI network with 21 nodes and 41 edges for down-regulated genes. Up-regulated genes are marked in green; down-regulated genes are marked in red

MiRNA-hub gene regulatory network construction

The miRNet database and Cytoscape software were used to establish the miRNA-hub gene regulatory network of the hub genes. A miRNA-hub gene regulatory network containing 2222 nodes (miRNAs: 1952 and hub genes: 270) and 9722 edges was constructed (Fig. 5). tfrc that was modulated by 108 miRNAs (ex: hsa-mir-409-3p); top2a that was modulated by 90 miRNAs (ex: hsa-mir-320d); cdk1 that was modulated by 68 miRNAs (ex hsa-mir-301a-5p); cep55 that was modulated by 56 miRNAs (ex: hsa-mir-107); fkbp5 that was modulated by 32 miRNAs (ex: hsa-mir-205-5p); atn1 that was modulated by 74 miRNAs (ex: hsa-mir-3200-3p); fscn1 that was modulated by 62 miRNAs (ex: hsa-mir-29a-5p); cherp that was modulated by 52 miRNAs (ex: hsa-mir-296-5p); notch1 that was modulated by 35 miRNAs (ex: hsa-mir-139-5p); notch3 that was modulated by 33 miRNAs (ex: hsa-mir-147a) (Table 5).

Fig. 5

Target gene—miRNA regulatory network between target genes. The blue color diamond nodes represent the key miRNAs; up-regulated genes are marked in green; down-regulated genes are marked in red

Table 5 miRNA—target gene and TF—target gene interaction

TF-hub gene regulatory network construction

The NetworkAnalyst database and Cytoscape software were used to establish the TF-hub gene regulatory network of the hub genes. A TF-hub gene regulatory network containing 356 nodes (TF: 78 and hub gene: 278) and 2152 edges was constructed (Fig. 6). rps27 that was modulated by 18 TFs (ex; creb1); fkbp5 that was modulated by 17 TFs (ex; tp63); tfrc that was modulated by 15 TFs (ex; foxa1); cdk1 that was modulated by 14 TFs (ex; nrf1); prdx1 that was modulated by 11 TFs (ex; jund); atn1 that was modulated by 19 TFs (ex; tead1); fscn1 that was modulated by 8 TFs (ex; srebf2); atp6v1b1 that was modulated by 7 TFs (ex; gata2); gata1 that was modulated by 7 TFs (ex; e2f1); mrpl12 that was modulated by 7 TFs (ex; tp53) (Table 5).

Fig. 6

Target gene—TF regulatory network between target genes. The gray color triangle nodes represent the key TFs; up-regulated genes are marked in green; down-regulated genes are marked in red

Validation of hub genes by receiver operating characteristic curve (ROC) analysis

To determine top hub genes have the diagnose significance of MI patients; The ROC analyses were conducted to explore the sensitivity and specificity of hub genes for MI diagnosis. The results showed that cftr (AUC = 0.895), cdk1 (AUC = 0.915), rps13 (AUC = 0.962), rps15a (AUC = 0.959), rps27 (AUC = 0.974), notch1 (AUC = 0.969), mrpl12 (AUC = 0.880), nos2 (AUC = 0.960), ccdc85b (AUC = 0.933) and atn1 (AUC = 0.943) have the best diagnostic value for differentiating the patients with MI from normal controls (Fig. 7). This indicated that expression of hub genes include cftr, cdk1, rps13, rps15a, rps27, notch1, mrpl12, nos2, ccdc85b and atn1 correlated with disease activity of MI, these hub genes could act as a biomarkers to estimate the activity of MI and verify the effectiveness of the treatment of MI.

Fig. 7

ROC curve analyses of hub genes. A CFTR, B CDK1, C RPS13, D RPS15A, E RPS27, F NOTCH1, G MRPL12, H NOS2, I CCDC85B, J ATN1

Discussion

Disregarding of new development in the treatment of MI, it has continued the greater frequent cause of heart-related deaths in the past few years. The huge mortality rate of MI is slightly due to the inadequacy of acceptable screening methods with tremendous sensitivity and specificity. Therefore, it is essential to determine potential biomarkers for screening and early diagnosis of MI. NGS technology has become essential tools for providing complete genetic information on MI samples and prophetic the changes in disease advancement.

The present investigation obtained NGS dataset GSE132143 from the GEO database, and total of 958 DEGs were screened, including 480 up-regulated genes and 478 down-regulated genes were selected between the MI samples and normal control samples. Research have shown that genes include il1rl1 [40] and alox15b [41] plays an important role in the pathogenesis of coronary artery disease. Some studies have shown that altered expression of genes include serpina3 [42], gpr78 [43] and esm1 [44] promotes the MI. A study indicates that scgn has been identified in diabetes mellitus [45]. These genes served as potential biomarkers for MI diagnosis and prognosis.

In this investigation, we identified enriched genes in GO terms and signaling pathways that might be utilized as diagnostic and/or therapeutic targets in MI. GO terms and signaling pathways include immune system [46], innate immune system [47], neutrophil degranulation [48], toll-like receptor cascades [49], metabolism of amino acids and derivatives [50], neuronal system [51], extracellular matrix organization [52], degradation of the extracellular matrix [53], diseases of glycosylation [54], response to stimulus [55], cell communication [56], cell periphery [57], membrane [58], anatomical structure development [59] and plasma membrane [60] were responsible for progression of MI. Recently, mounting researches have revealed that genes include pla2g2a [61], ccl23 [62], cd53 [63], treml4 [64], trem2 [65], cd180 [66], hpse [67], cela2a [68], tnfrsf4 [69], ambp [70], sox18 [71], panx2 [72], rspo2 [73], comp [74], asgr1 [75] and noxa1 [76] were vital for the onset and developmental process of atherosclerosis. A great number of studies have indicated that genes include s100a9 [77], adora3 [78], il1r2 [79], fpr1 [80], ccl20 [81], cd163 [82], s100a8 [83], tlr2 [84], has2 [85], ptx3 [86], timp4 [87], areg [88], lbp [89], il18r1 [90], alox5ap [91], retn [92], f13a1 [93], fpr2 [94], saa1 [95], flt3 [96], aqp4 [97], fcer1g [98], ccl18 [99], hp [100], cdk1 [101], slc7a11 [102], cftr [103], f8 [104], stc1 [44], il18rap [90], timp3 [105], pde4d [106], cyp4a11 [107], scn10a [108], apob [109], ace [110], penk [111], hspb6 [112], tlr9 [113], egr1 [114], cacng8 [115], foxd3 [116], dbh [117], foxp3 [118], glp1r [119], il34 [120], ccn1 [121], adra2a [122], bgn [123], nos2 [124], agrn [125], drd1 [126], gnb3 [127], egr2 [128], mdk [129], notch3 [130], azin2 [131], notch1 [132], loxl2 [133], adamts14 [134] and sod3 [135] have been implicated in MI pathology.. A previous bioinformatics study suggested that genes include defb1 [136], slc11a1 [137], spink1 [138], ccr1 [139], glul [140], gpr84 [141], siglec5 [142], siglec7 [143], lgr5 [144], cd38 [145], grb14 [146], prdx1 [147], slc19a2 [148], cadm2 [149], trpm5 [150], col1a1 [151], ctrb1 [152], uts2r [153], crtc1 [154], muc5b [155], tmprss6 [156], slc5a2 [157] and kcnj9 [158] might play a role in the development of diabetes mellitus. A previous study reported that the genes include mt1a [159], lyve1 [160], s100a12 [161], gckr [162], tlr8 [163], mrc1 [164], agtr1 [165], p2ry12 [166], msr1 [167], nqo1 [168], fkbp5 [169], cmtm5 [170], adh1c [171], aplnr [172], sfrp4 [173], ccl3 [174], col11a2 [175], egr3 [176] and il2rb [177] play an important role in the pathophysiology of coronary artery disease. Study demonstrated that genes include edn2 [178], snx10 [179] and kcnn1 [180] can participate in the occurrence and development of atrial fibrillation. The abnormal expression of genes include ednrb [181], fgf10 [182], wnk3 [183], kng1 [184], fcn3 [185], aqp3 [186], hpr [187], cth [188], sparcl1 [189], maoa [190], bmpr1b [191], fgf7 [192], calcrl [193], mark3 [194], adh1b [195], amh [196], ret [197], igf2 [198], slc6a9 [199], nppa [200], sct [201], dcx [202], asic1 [203], lmx1b [204], dbp [205] and slc6a9 [206] contributes to the progression of hypertension. Study showed that the genes include camp [207], spp1 [208]. cxcl11 [209], tfrc [210], irak3 [211], c3ar1 [212], gli2 [213], thy1 [214], foxo6 [215], rgs4 [216], wnt10b [217] and aebp1 [218] plays a vital role in the development of might be related to the pathophysiology of obesity. A study showed genes include tdgf1 [219], kif20a [220] and ltbp2 [221] are highly prone to congenital heart defect. Our data indicated that enriched genes might have impact on MI progression. It would be meaningful to confirm exact function of enriched genes in MI in prospective.

Building PPI network is favorable for investigators to study the underlying molecular mechanism of MI for the reason that the DEGs would be unified and organized in the network deciding by their interactions. Identification of hub genes that might be key prognostic markers, diagnostic markers and therapeutic targets for MI. Based on the PPI network constructed by the online database HIPIE, we identified hub genes. A recent study suggested that the hub genes include cftr [103], cdk1 [101], notch1 [132], nos2 [124] and notch3 [130] might take part in the progression of MI. We gave a new confirmation for that rps13, rps15a, rps27, rpl26, rps29, mrpl12, ccdc85b and atn1 are expected to become a novel biomarker for MI prognosis. Thus, it might consider as potential therapeutic targets for MI.

To illuminate the potential molecular mechanism of the hub genes in MI, we focused a miRNA-hub gene regulatory network and TF-hub gene regulatory network analysis. The identified hub gens, TFs and miRNAs might be associated in the pathological process of MI. Zaja et al. [101], Si et al. [132], Zhang et al. [130], Chen et al. [222], Li and Zhang [223], Wang et al. [224], Zhao et al. [225], Lin et al. [226], Liao et al. [227], Izadpanah et al. [228], Wang et al. [229] and Hakobjanyan et al. [230] reported that the expression of the cdk1, notch1, notch3, hsa-mir-409-3p, hsa-mir-320d, hsa-mir-107, hsa-mir-139-5p, nrf1, tead1, gata2, e2f1 and tp53 can lead to MI. Studies reported that biomarkers include hsa-mir-301a-5p [231], hsa-mir-29a-5p [232], hsa-mir-296-5p [233], creb1 [234] and foxa1 [235] were proposed to contribute to the development of diabetes mellitus. A previous study reported that hsa-mir-205-5p plays a key role in hypertension [236]. Accumulated evidence has demonstrated that biomarkers include fkbp5 [169] and srebf2 [237] are associated with coronary artery disease. Studies have shown that biomarker tfrc (transferrin receptor) [210] was identified to be associated with obesity. Our interesting finding is that novel biomarkers include top2a, cep55, fscn1, cherp, atp6v1b1, gata1, hsa-mir-3200-3p, hsa-mir-147a, tp63 and jund (jun D proto-oncogene) are all closely related to MI. Evidence from studies indicates that these novel biomarkers might be related to the advancement of MI and thus has the potential to be used as a diagnostic biomarkers of MI.

Our investigations had some limitations. Firstly, we used only 1 dataset, and there might have been bias in the search for biomarkers for MI. Secondly, we did not account for the potential confounding effects of demographic variables. Finally, we did not carry out experiments to prove the research results, which will be considered in future work. Finally, we did not carry out in vivo and in vitro validation experiments for hub genes to prove the research results, which will be considered in future work.

Conclusion

This investigation identified significant DEGs between MI and normal control samples via analyzing NGS dataset. cftr, cdk1, rps13, rps15a, rps27, notch1, mrpl12, nos2, ccdc85b and atn1 were verified and considered as hub genes were linked with disease prognosis, which could be predictive and therapeutic targets. The above results provide novel investigation directions for the relationship between MI and its associated complications. Bioinformatics and NGS technology provide opportunities to further examine the molecular mechanism and targeted therapy of MI on a transcriptional level. In vivo and in vitro validation experiments will be implemented in ongoing work.