Gene Networks Reveal LASP1, TUBA1C, and S100A6 as Likely Drivers of Multiple Sclerosis (MS)


 Background: Multiple sclerosis (MS), a non-contagious and chronic disease of the central nervous system, is an unpredictable and indirectly inherited disease affecting different people in different ways.Using Omics platforms data i.e. genomics, transcriptomics, proteomics, epigenomics, interactomics, and metabolomics it is now possible to construct sound systems biology models to extract full knowledge of the MS and path the way to likely construct personalized and therapeutic tools.Methods: In this study, we learned many Bayesian Networks in order to find the transcriptional gene regulation networks that drive MS disease. We used a set of BN algorithms using the R add-on package bnlearn. The BN results underwent further downstream analysis and were validated using a wide range of Cytoscape algorithms, web based computational tools and qPCR amplification of blood samples from 56 MS patients and 44 healthy controls. The results were semantically integrated to come up with improved understanding of the complex molecular architecture underlying MS, distinguishing distinct metabolic pathways and providing a valuable foundation for the discovery of deriven genes and possibly new treatments. Results: Results show that the LASP1, TUBA1C, and S100A6 genes were most significant and likely playing a biological role in MS development. Results from qPCR showed a significant increase (P <0.05) in LASP1 and S100A6 gene expression levels in MS patients compared to controls. However, a significant down regulation of TUBA1C was observed in the same comparison. Conclusion: This study provides potential diagnostic and therapeutic biomarkers for increasing understanding of gene regulation underlying MS.


Introduction
Miller DH, 2008 #106;Miller DH, 2008 #106}Multiple sclerosis (MS) is a multifocal in ammatory autoimmune disease that severely affects the white matter of the CNS and leads to progressive neuronal damage in genetically sensitive hosts [1]. The integration of information gleaned from a variety of resources e.g. transcriptomics, genomics, proteomics and patient clinical data could boost our understanding of the mechanism(s) underpinning this disease [2]. In this regard, we can explore the signaling pathways involved in MS [3], apply logical networks to model signaling pathways in MS [4] and use networks to combine information on transcriptome-interactome data from MS studies [5]. We can also apply theory of biochemical systems for improving therapeutic drugs in re-myelination [6], create molecular networks based on transcription factors and genes expressed in mononuclear cells in MS patients [7], and design reactive networks between distinct miRNA and target genes in T cells [7] -all of which will help explain the molecular mechanisms involved in MS disease [8]. Additional Table 1 shows some examples of network based studies used with different MS biological data. As one can see, Bayesian Network (BN) modeling paradigms have rarely been applied in this setting. BN uses probability theory to reason under uncertainty. BN as a graphical scheme (directed acyclic graph) consists of a qualitative part (structural model) and a quantitative part (local probability distributions), which allows for a different kind of probabilistic inference, and quantitatively measures even the smallest impact of a variable or set of variables on others. This sort of modeling is of great importance in transcriptomic studies, since it can reveal both qualitative and quantitative elements of learned gene networks. BN has previously been used in several transcriptomic studies [9][10][11].
Many existing categories of gene networks identify groups of related genes as gene sets, making experimental follow-up a formidable task. With BN, it is possible to determine whether a gene is a driving source of changes in its gene network or not, since both in-degree and out-degrees of connectivity of each gene can be readily veri ed. The more out-degree each gene has, the higher likelihood it has of being a key regulator gene. This would be a crucial characteristic when looking for potential drug targets. Today, MS research is increasingly data-driven -a trend that arguably shall continue at a much higher rate in times to come. To tackle these large amounts of heterogeneous data, and to derive insight into MS disease, many interdisciplinary scientists have turned to a variety of computational tools. In this study we aim to gain much better insight into the regulatory transcriptional gene network underlying MS using systems biology approaches in the context of BN, that may yield mechanistically interpretable results.

Network analysis
In this study the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/ ) was scanned using a combination of several simple key words, and resulting DNA microarray experiments related to MS, that ful lled our criteria, were explored. In the end, based on our criteria for choosing a suitable GEO data set, the microarray series with accession number GSE17048 was downloaded from GEO using the GEOquery package [12]. This accession was seen to have the highest number of arrays per probe -a fact that would help minimize the rate of false positives while training the regulatory gene BN. The GSE17048 study represents mRNA expression for all known genes in whole blood from 144 individuals (99 MS patients and 45 age-matched controls). In order to remove noise from the data, probes with the highest variance were obtained and used as an input to train the gene regulatory BN using bnlearn, an R add-on package [13,14]. The following codes were used to lter the probes with highest variances: qt <-quantile( t(data1) ; probs = c(0.0002,0.99) ); rows1 <-apply( t(data1) , function(x) any( x < qt | x > qt) ); data2 <-t(data1) [, rows1] We obtained the best tted BN model on our data using Bayesian information criterion (BIC) and its adjacency matrix, with the help of the Cytoscape-based aMatReader plugin, with the Cytoscape [15] environment used for further downstream scrutiny.

Downstream analysis
This was accomplished with the following Cytoscape add-on packages: the jActiveModules to explore the concept of gene modules and nd sub-networks [16]; MCODE to identify putative complexes by nding regions of signi cant local density [17]; CytoHubba, to explore the PPI network of hub genes using eleven different methods [18]; the NetworkAnalyzer to determine the hub genes, taking into account the degree of topological criteria (e.g. the number of nodes, edges, and connected components, along with the network diameter, radius, density, centralization, heterogeneity, and clustering coe cient., The characteristic path length, and the distributions of node degrees, neighborhood connectivities, average clustering coe cients, and shortest path lengths were also considered) [19]; iRegulon to detect targets / motifs / paths from a set of genes; and the CyTargetLinker to integrate regulatory reactions in network analysis. In addition, we used Metascape [20] to annotate the multiple gene lists from our study. Even though transcriptomic statistical analysis is generally based on probe level data, the probe names were converted to their corresponding gene names using g:Pro ler (https://biit.cs.ut.ee/gpro ler/gost) to get better insight into the data. Results from the aforementioned software were combined. Figure 1 shows the analysis pipeline used in this study.
Validation of LASP1, TUBA1C, and S100A6 genes using Quantitative Real-Time PCR (QRT-PCR) Following up on outputs from the bioinformatics analysis, validation of signi cant DEGs (LASP1, TUBA1C and S100A6) was carried out using QRT-PCR. A total of 100 whole blood samples (56 MS cases, mean age: 39.5 years and 44 controls, mean age: 39.5) were provided by Al-Zahra Hospital and a Multiple Sclerosis medical center in Isfahan province. Within the 56 MS cases, 44 patients were in RRMS (Relapsing-remitting MS) phase and 12 patients were in SPMS (Secondary progressive MS) phase.

RNA extraction
Total RNA was extracted from each sample according to the standard TRIzol protocol (Bio BASIC, Canada) according to manufacturer's instructions. RNA concentration and quality were determined using both Nanodrop (Thermo Scienti c Tm Nano Drope One C model) and gel electrophoresis -existence of two sharp bands representing 18S and 28S ribosomal RNA on a 1% (w/v) ethidium bromide stained agarose gel during electrophoresis through TAE buffer (40 mM Tris-acetate, 1 mM EDTA, pH 8.0) at 100 V for 30 min. Those RNA samples with a RNA integrity Number (RIN) < 1.8 were excluded from further analysis. High quality extracted total RNA was stored at -70°C until cDNA synthesis. cDNA synthesis Initially, DNAse I (Fermentase Cat # ENO 521) treatment was used to remove genomic DNA before cDNA synthesis. Next, cDNA synthesis was carried out using a commercial kit provided by Yektatajhiz Company (Cat No:YT4500) according to manufacturer's instructions. e.g. keeping samples on ice under sterile conditions at 70°C for 5 mins , 37°C for 60 mins, 70°C for 5 mins, and nally storing all synthesized cDNAs at -20°C.

Quantitative Real Time PCR Analysis
To enable the validation of our candidate genes (LASP1, TUBA1C and S100A6), SYBR Green -based QRT-PCR was performed using a LightCycler® 96 (BioRad, Germany). The sequence of all primers used are listed in Table 1. These were designed using the PRIMER3 program (http:// frodo.wi.mit.edu). QRT-PCR reactions were performed in duplicate and the values of average cycle threshold (Ct) were determined for each sample. The conditions of QRT-PCR ampli cation were: 1 cycle at 95 °C for 2 min, 40 cycles at 95 °C for 50 s, 60 °C for 30 s. The human beta-actin gene (ACTB_HUMAN) was used as the internal control. Hence, all calculated concentrations are relative to the concentration of the standard, expressed in arbitrary units and the quanti cation cycle (Cq) values were automatically calculated with Rotor-Gene software version 6.1. The results were analyzed using the 2 −ΔΔCt method [21].

Results
The fundamental idea behind this analysis was to shed some light into gene-gene interactions underpinning MS disease with regard to cause and effect. The bnlearn package was used to nd the gene regulatory network behind this disease [22]. The results of comparison of the network structures determined from various algorithms, including Hill Climbing, Tabu Search, Max-Min Hill Climbing, and Restricted Maximize algorithms with different scoring functions, are shown in Table 2. Some key properties of BN are fundamental in judging estimated results. Topological parameters can characterize the location of genes in a gene network [23]. Using NetworkAnalyzer the following network topological parameters were calculated in our data: clustering coe cient (0.003), number of nodes (1707), connected components (857), network diameter (27), network radius (1), shortest paths (192463), characteristics path (8.313), the average number of neighbors (1.992), network density (0.0), isolated nodes (854), number of self-loops (0), multi-edge node pairs (0) and analysis times (1.467). The nature of probability distribution induced by a gene regulatory BN will allow diverse probabilistic gene queries to be answered in linear time. However, many structural BN parameters may be important. One of the key parameters in Table 2 is the branching factor. This parameter plays a signi cant role in development of the gene network. Each Node (gene) will have its own branching measure, which will determine the outdegree of that gene. If the branching factor value is not uniform in the network, an average branching factor can generally be calculated. This value turned out to be different depending on which algorithm was used. Max-Min Hill Climbing returned a higher average than Restricted Maximize. In terms of system level understanding of research, the higher the branching factor, the more frequently gene regulators can be identi ed in the network.
Biological networks have a modular architecture [24]. MCODE can nd connected and dense areas of the gene network based on network topology measures. In our analysis, 12 different modules were detected using MCODE, among which, 7 modules had 3 nodes; 3 edges with different interaction modes; 3 modules had 6 nodes and 7 edges; 1 module 15 nodes and 17 edges and nally 1 module had 6 nodes and 6 edges (Additional Fig 1). The active subnetworks were obtained using jActiveModules. The jActiveModules created 5 modules, in which ILLMN_1742167 (TUBA1C), ILLMN_1665909 (LASP1) and ILLMN_1713636 (S100A6) were seen to be enriched in those modules. (Figure 3 and Additional Fig 2). The number of modules detected by this method was different than those identi ed with the MCODE based method. Figure 3 shows predicted modules in different modes of interaction. Module-level analysis explores the organization of biological systems and reconstructs module networks. Figure 3 shows a module-level view of our gene regulatory BN network that denotes a high level representation of the regulatory machinery of the MS gene network topology. Dense module searching of two MS GWAS datasets identi ed several genes (GRB2, HDAC1, IL2RA, JAK2, KEAP1, MAPK1, RELA and STAT3). These genes are enriched for glial cell differentiation [25]. CytoHubba provides a user-friendly interface for discovering important nodes in biological networks [26]. Cytohubba considers the shortest path between groups of nodes. Among the 11 proposed algorithms, MCC tted better than the others. In Figure 4 and Table 3 we present the top 10 identi ed probes.
Many genes e.g. TUBA1C, LASP1 and S100A6 indicated in Figure 4 are close to being hub genes, and were indeed identi ed as hub genes by other algorithms. CyTargetLinker sheds some light into the network in terms of regulatory interactions i.e. miRNA-target, TF-target or drug-target interactions. The ErbB Signaling pathway for MS disease is available in the WikiPathways database and the network created for ErbB is shown in Additional Fig 3. The iRegulon software then allowed us to identify regulons using motif discovery in a set of regulated genes. Identi ed transcription factors affecting the hub genes are listed in Additional Table 2 Table 4. The most signi cant, the STAT5A protein, mediates the responses of many cell ligands, such as IL2, IL3 and different growth hormones.
In this study, the gene identi ers were uploaded to Metascape and used in conjunction with KEGG pathways, GO biological processes, Reactome gene complexes, canonical and CORUM pathways [27]. The results of the enrichment analysis, including descriptions, function, ontology, expression, etc. are shown in Table 5, Additional Table 3 and Additional Fig 5. Genes were ranked from top to bottom based on degree, closeness and betweenness [higher degree (hub), higher betweenness (throat) and higher closeness centrality (shortest distance with other genes in the network]. In terms of these parameters, three genes (LASP1, TUBA1C, S100A6 ) showed a signi cant correlation with MS disease. These were thus identi ed as hub genes (Additional Table 4). In this study, probes ILLMN_1665909, ILLMN_1742167, ILLMN_1713636 had high degrees of 15, 13 and 11, respectively and were identi ed as hub probes. 850 probes had zero input edges, 200 probes had 1 in-degree. ILLMN_1665909, with the highest out-degree (13 out-degree) and 2 in-degree (mapped to human LASP1) plays an important role in regulating activity. Its encoded protein is a cytoplasmic protein that binds focal adhesion proteins and plays a role in cell signaling, migration, and proliferation. ILLMN_1742167, with 12 out-degree and 1 in-degree mapped to the human tubulin gene (TUBA1C), and ILLMN_1713636 with 9 out-degree and 2 in-degree mapped to the S100A6 gene. (Figure 2 and Additional Fig 6).
Patterns of relative gene expression for these genes (LASP1, TUBA1C and S100A6) showed signi cance between MS cases and controls (P <0.05). Our results demonstrate that LASP1 and S100A6 genes show a signi cant up-regulated gene expression pattern while TUBA1C is signi cantly down-regulated when MS cases are compared to healthy controls (Table 6 and Figure 5).
The results after studying the normality of the distribution of variables using the one sample Kolmogorov smirnov test and Unpaired t-test in GraphPad Prism 8 software show a signi cant difference in expression levels of LASP1, TUBA1C and S100A6 genes between patients and healthy controls. P-values were: TUBA1C <0.0001, S100A6 <0.0001, LASP1 <0.003. Mean expression of TUBA1C, LASP1 and S100A6 genes in patient samples was 7.4, 5.6 and 2.9 respectively and 4.9, 8.1 and8.0, respectively in healthy individuals. Results from statistical analysis also show a decrease in TUBA1C gene expression and an increase in LASP1 and S100A6 gene expression in MS patients compared to the control group ( Figure 6).

Discussion
At present, the cause of MS is not fully understood, but knowledge of the genetic factors involved is essential for effective diagnosis and identi cation of the most appropriate MS therapeutic interventions. In this study, three genes (LASP1, TUBA1c and S100A6) with high degree, high closeness centrality and high betweenness measures, were highlighted as potential MS candidate regulator markers. These three genes (LASP1, TUBA1c and S100A6) seem to be the most signi cant genes in the MS disease process. Patients with MS are known to suffer from a number of digestive problems [28] and studies have shown that LASP1 [29] and S100A6 genes have high expression in the digestive system. A link can therefore be established between the expression of these genes, MS and gastrointestinal problems in MS patients. S100A6 functioning in a wide range of cell types, is a member of the S100 family and may be involved in and expressed in several types of human cancers [30]. S100 family expression in MS patients could be considered as a diagnostic biomarker for MS, in which its inhibition of demyelinating nerve cells shows that S100 proteins could act as a candidate therapeutic target in MS [31]. Keiko Komatsu et al.(2000), reported increased expression of S100A6 (Calcyclin), a calcium-bound protein of the S100 family, in human colorectal adenocarcinoma [32]; in this way, Eva Peterova (2021) reported an overexpression of S100 protein-encoding mRNA in both colorectal cancer cell lines [33]. A study by Bartkowska et al. [34] showed that in response to different stress conditions, the level of S100A6 decreased in several brain structures, indicating that S100A6 may modulate stress responses. In an autoimmune disease study, a genomewide methylation array has identi ed a few hypomethylated immune-related genes, amongst them S100A6 which shows up-regulation in autoimmune encephalitis patients [35].
Even though S100A6 is involved in many biological phenomena e.g. cell proliferation, cytoskeletal dynamics and tumorigenesis, and some of its biological activity is still unknown [36]. At the transcriptional level, upstream stimulatory factor and Nuclear factor-kappa B (NF-κB) activates the S100A6 gene promoter, although p53 might act indirectly to suppress transcription of the S100A6 gene [37].
Microtubules, that are essential for multiple cellular processes, are constructed from homologous tubulin proteins. Microtubules shape crucial cytoskeletal structures that play a pivotal role in creating and maintaining neuronal mechanistic, biochemical functions such as polarity, regulating neuronal morphology, dendrite growth, neuron migration throughout brain development, transporting cargo, and scaffolding signaling molecules to form signaling hubs, and are vulnerable to degradation and disorganization in a variety of neurodegenerative diseases [38][39][40]. Malfunction of microtubules (Tuba1c) is also considered as the central physiopathological mechanism of neurodegenerative diseases and abnormalities in the regulatory pathways of microtubules disrupt the properties and functions of microtubules, leading to nerve damage [41]. In a 2006 study, a decreased expression of the TUBA1C gene in Parkinson's disease was demonstrated by quantitative analysis of gene expression [42]. LASP1 (The LIM and SH3 protein 1) a focal adhesion adaptor protein, is an actin-binding, signaling pathway-regulated phosphoprotein and localizes within multiple sites of dynamic actin assembly. It has the potential to interact with various molecules, and is highly expressed in the adult central nervous system. It is ubiquitously expressed in normal tissues, and it transmits signals from the cytoplasm into the nucleus. It is a versatile structural, signaling, and biomarker protein and plays a crucial role in the growth and metastasis of gastric cancer, as well as being overexpressed in several other cancers [43][44][45][46][47][48]. Also, overexpression of LASP1 is often seen in colon cancer tissues, especially in metastatic colon cancer tissues. LASP1 can cause the progression and metastasis of colorectal cancer (CRC), but its mechanism is still unclear [49]. Another study showed that LASP1 binds to the calcium-binding protein family (S100A) and increases its expression in colon cancer [50]. this way, statistical analysis revealed that LASP1 and S100P are correlated (Kappa = 0.347, P < 0.01). Microarray data has revealed that alterations in LASP1 proteins affect cell migration, adhesion, and cytoskeletal organization [47]. LASP1, signi cantly expressed by CNS neurons, is localized at synaptic sites [51].
A couple of signi cant transcription factors that interact with these hub genes were identi ed in this study. The YY1 transcription factor (Yi and Yang 1) is a multifactorial protein that, depending on the cell tissue, can activate or suppress gene expression [52]. It is expressed in the nervous system. The YY1 promoter lacks the usual TATA box but has a rich GC sequence and therefore resembles a large subset of housekeeping and growth regulator genes. These features suggest that it may play an important role in development. In the central nervous system, myelination is performed by oligodendrocytes. YY1 function in oligodendrocytes was rst reported by Berndt et al (2001). YY1 activates the promoter of myelin lipids and has been identi ed as an important player in myelination of the central nervous system during growth. In multiple neurodegenerative diseases, YY1 function is degraded through distinct mechanisms, including protein utilization, protein degradation, and ectopic nuclear / cytoplasmic shuttle (N/C). These disorders inhibit YY1 transcriptional activity and lead to gene transcriptional abnormalities that contribute to disease pathogenesis. A future goal in YY1 research is to discover other potential mechanisms that lead to YY1 dysfunction in neurodegenerative diseases, such as ectopic changes after translation [53].
NFAT is a family of transcription factors originally described as key factors in the immune response. There are ve major members of the NFAT family: NFATc3 is one of ve members of the NFAT transcription factor family that act as signal integrators because their function is to bind STAT3, c-Jun, CREB, and ATF3 factors at speci c DNA binding sites. NFATs cannot be regulated alone and act as calcium-dependent transcription factors. The antigen-mediated T cell receptor (TCR) mediates multiple signaling cascades, including phospholipase C (PLC) -dependent pathways that are secondary messengers of inositol-1,4,5-triphosphate (IP3) and diacylglycerol (DAG). IP3 binds to the IP3 receptor in the endoplasmic reticulum (ER) and releases Ca 2+ ions into the cytoplasm. [54].
NFATc1-4 activates intracellular calcium via dephosphorylation. Calcinurin activates a calcium / calmodulin-dependent phosphatase, which detects the transmitter signal to the nucleus (NLS) and carries it to the nucleus. This acts as a link between calcium signaling and NFAT-dependent gene transcription. Conversely, NFAT inactivation is mediated by phosphorylation by kinases that either retain NFAT in the cytoplasm or enhance nuclear export. According to a recent study, native human astrocytes also express NFAT and transmit to the nucleus [21]. The ndings suggest that NFATs control pathways involved in astrocyte activation and, therefore, may affect neuronal cell survival. Such a compelling idea suggests that NFAT is employed in neurological pathologies associated with neuro-in ammation in astrocytes. The ndings show that NFATc3 is de ned as a marker of a speci c subset of astrocytes that are activated in response to lesions, as well as some degree of heterogeneity among astrocytes that may have consequences for cells in the nervous system [55]. Preliminary ndings in neuroblast cells have shown that various treatments that alter tubulin polymerization, such as reducing the mineral zinc, prevent the transfer of NFAT to the nucleus.
In addition, NFAT clusters gradually move away from the microtubules and communicate by transporting NFAT to the nucleus. Therefore, it has been shown that the association of NFAT with the microtubule network can increase the concentration of this transcription factor around the nucleus, and/or facilitate its interaction with nuclear pores. In agreement with a functional relationship between NFAT and microtubules, it has been observed that the degradation of several proteins that control the proper organization of the microtubule network, and the actin-cytoskeletal linker, disrupts the nucleus and transcriptional activity of NFAT. Overall, it indicates the involvement of microtubules in NFAT nuclear stimulation [56]. The LASP1 gene enhances NFAT2 nuclear translocation by activating the nuclear factor Akt [57]. NFAT can affect processes such as axon growth, synaptogenesis, Schwann cell differentiation, and myelination [55]. In general, it can be concluded that increase of the expression of LASP1 and S100A6 genes and decrease the expression of the TUBA1C gene in multiple sclerosis disrupts NFAT transcriptional activity. Although the role of NFAT in regulating the immune system is well established, our knowledge of NFAT in human disease is limited. The function of NFAT in other aspects of human immune or in ammatory diseases is also largely unknown [58].
On this basis, the present study con rmed the importance of three gene expression patterns (LASP1, S100A6, TUBA1C) for understanding the transcriptome complexity of MS. This leads us to conclude that upregulation of LASP1 and S100A6 genes along with down-regulation of TUBA1C is central to MS pathology. To our knowledge, this is the rst report to evaluate the level of expression of the above genes for discovery of a transcriptomic signature for MS disease. These ndings provide a potential mechanism for some signi cant biomarkers responsible for the pathogenesis of MS. However, we still have a long way to go to understanding the larger transcriptomic pro le for this disease. This study provides initial data to further investigate the possible role of these genes in the pathogenesis of MS.

Conclusions
NetworkAnalyzer is a versatile tool for the analysis of biological and other networks. In this study, we implement our network scoring methods into a Cytoscape plugin, cytoHubba. Results of the present study indicate that the analysis of gene expression data based on gene-gene interaction networks can provide more opportunities to determine the genes involved in MS. The importance of three candidate marker genes in this disease were highlighted -LASP1, TUBA1C and S100A6. These candidate genes identi ed by the biological systems approach, have been further con rmed in the laboratory. The signi cant difference in the expression of these three genes in patients with MS can be concluded to have a signi cant effect during this disease and will help further research on MS and its treatment. This useful tool can serve as a good starting point for identifying new therapies and understanding the basic mechanisms controlling normal cellular processes and disease pathologies. All experimental procedures were approved by the Ethics Committees of the University of Isfahan, and informed consent was obtained from all participant prior to in the study.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed in the current study are available from the corresponding author on reasonable request.  loglik-g: The multivariate Gaussian log-likelihood (loglik-g) score, AIC-g: Akaike Information Criterion score, BIC-g: Bayesian Information Criterion score.   group) by a mean factor of 5.491 (S.E. range is 1.183 -32.843); LASP1 sample group is different to control group. P(H1)=0.000. TUBA1c is DOWN-regulated in sample group (in comparison to control group) by a mean factor of 0.166 (S.E. range is 0.058 -0.523). TUBA1c sample group is different to control group. P(H1)=0.000. S100a6 is UP-regulated in sample group (in comparison to control group) by a mean factor of 36.556 (S.E. range is 9.630 -140.562).S100A6 sample group is different to control group. P(H1)=0.000 Figure 1 The analysis pipeline used in this study.

Figure 2
Some topological measures of trained gene regulatory BN visualized by NetworkAnalyzer.

Figure 4
The MCC method captures essential genes in the top ranked list. Figure 5