Graph theoretical way of understanding protein-protein interaction in ovarian cancer

Abstract

Cancer is a major research area in the medical field. Precise assessment of non-similar cancer types holds great significance in according to better treatment and reducing the risk of destructiveness in patients’ health. Cancer comprises a ambient that differs in response to therapy, signaling mechanisms, cytology and physiology. Netting theory and graph theory jointly gives a viable way to probe the proteomic specific data of cancer types such as ovarian, colon, breast, oral, cervical, prostate, and lung. We observe that the P2P(protein-protein) interaction Nettings of the cancerous tissues blended with the seven cancers and normal have same structural attributes. But some of these point to desultory changes from the disease Nettings to normal implying the variation in the dealings and bring out the redoing in the complicacy of various cancers. The Netting-based approach has a pertinent role in precision oncology. Cancer can be better dealt with through mutated pathways or Nettings in preference to individual mutations and that the utility value of repositioned drugs can be understood from disease modules in molecular Nettings. In this paper, we demonstrate how the graph theory and neural Nettings act as vital tools for understanding cancer and other types such as ovarian cancer at the zeroth level.

Keywords

Cancer ovarian cancer graph parameters protein nettings fuzzy neural nettings

1 Introduction

The transformation of normal cells into abnormally proliferating cancer cells is attributed to the changes in the genes or proteins and their altered functions in the altered signaling pathways. The gene/protein modifications can be genetic and epigenetic which can activate or suppress the expression of genes/proteins in healthy cells. The advancements in whole genome sequencing have enabled the identification of tumor suppressor genes and oncogenes. Often cancer is observed as a condition with modified cellular pathways owing to the mutations of genes/proteins which can be common or distinct in the tumors of same type. Cancer is not a single gene disorder and the mechanisms unraveled so far as to understand the intricacies in the complex Netting is due to the research enhancements in system biology. This has led to the discovery of a collection of Netting-based computational methods to identify cancer pathways through different genomic data.

Netting dependent probe paved to attention in the repositioning of the drug to lower the cost of new drug design by making use of existing drugs on unique targets in drug-target interaction Nettings. Various Netting-dependent ways assisted to learn Nettings of drugs, targets, and diseases to reposition it for other new targets. It is high time to give a unified viewpoint of graph theory procedures for Netting analysis.

The circuit flow algorithm recognizes differentially expressed genes. Then it is applied to locate causal paths from the causal genes/proteins to the target genes/proteins in protein-protein interaction. Also, the causal genes are picked to describe every differentially expressed target genes/proteins. HotNetalters gene variations in a gene Netting and then employs a diffusion kernel to construct an interaction graph with the edges weighted by the interaction between each pair of genes. Combinatorial tasks are placed to discover the subNettings of modified genes in a significant sum of patients. NetPathIDemploys Netting diffusion to study 16 types of cancers. PARADIGM is a probability dependent graph model used to model the gene transcription events. Each gene is modeled by considering a factor graph of DNA copy numbers. The factor graphs of genes are linked depending on their regulatory relations in a pathway. The proteomic and genomic data are probed in the mathematical graph models for the implication of pathway activities in each patient. It leads to the derivation of integrated pathway activity (IPA) scores. Mutually exclusive mutated genes are placed on the same pathway, and Netting methods propagate the signals to identify the pathway by an appropriate signal. Co-expression of modified genes as well conjointly strengthens the mutation signals. This observation support that the patterns of mutation are accurately captured on ovarian cancer by label propagation.

To get a flavor of both theory and application it is recommended to see the main concepts of graph theory with the mathematical and empirical explanation of graphs that denote Nettings as in [12, 25]. A graph G can be visualized as a 2-tuple (V, E) where V is a nonempty collection of elements denoting the nodes and E is a collection of links denoting the join among the nodes. Mathematically E={(r, s)| r, s ∈ V} the lone link between vertices r and s. We call r and s are neighbors in this case. A multi-edge link comprises≥2 links with the same ends. These are pertinent for Nettings where two nodes can be linked by≥1 connections. In that case, each join carries distinct information. Such a feature can be seen in P2P relation Nettings. For a PPI database with such a feature, one can refer [28]. A directed graph also referred to as digraph is 3-tuple G=(V, E, g), where g associates to every member in E an structured pair of members in V. We call structured pairs of such members as arcs marked with direction. So E=(r, s) is considered as directed from r to s. Digraphs are appropriate to explain pathways that are biological which indicate the relation occurring in nature as a sequence at time points that are either single or many and the channel of counsel all through the netting. These are mostly regulatory nettting or metabolic or signal transduction [32]. A graph with weight is G=(V, E) where V is a non-empty collection of elements and E is a collection of links among the elements defined as E = {(u, v) | u, v ∈ V} with a real-valued map w: E ⟶ R. The weight w_rs of the link among r and s points to the pertinence of the link. Generally, a higher link weight indicates more solidity of a link. Graphs with weight are adopted netting in bioinformatics. For instance, relations whose priority alters are often marked to data that are biological to attain the importance of co-succulence located by mining of the text among co-expression of genes [28, 39]. By a bipartite graph G=(V, E) we mean a graph with no direction marked for arcs, where V is divided into 2 V₁, V₂ such that (u, v) ∈ E means either u ∈ V₁ and v ∈ V₂ or v ∈ V₁ and u ∈ V₂. These are used in the visualization of nettings that are biological ranging from denoting relation that is enzyme-reaction dependent [50] or [7]. G₁=(V₁, E₁) ⊆ G=(V, E) if V₁⊆ V and E₁⊆ E, where each link in E₁ join vertices in V₁. Degree of a node in a graph that has no direction for its arcs we deem the number of links the node has to other nodes and is set as |N(u)| = deg(u) where N(u) is the cardinality of N(u). Suppose the netting is marked with direction for arcs, then every node has two variations in-degree deg_in(u) (for the number of incoming links to node u), and the out-degree deg_out(u) (for the number of outgoing links from vertex u. A total connectivity of a netting is C=|E|/ n(n–1) where |E| is the cardinality of E and n is the cardinality of V. This structural parameter is very useful in a food-web Netting to represent reaction interplay and reversibility, This can be seen by using software like MixNet [45].

The adjacency matrix representation of a graph G=(V, E) comprise a n×n matrix A=(a_u,v) such that a_u,v=1 if (u, v) ∈ V and a_u,v=0 otherwise. In weighted graphs a_u,v=w_uv if (u, v) ∈ V or a_u,v=0 otherwise. The matrix is symmetric for undirected graphs as a_r,s=a_s,r. But it is not true for digraphs, as the lower and upper triangular portions of the array convey the link direction. These arrays require a Θ(|V|²) amount of space and are appropriate for dense graph like structure. This is suitable for clustered nettings, where the link density is high among elements. For a struture where each vertex is linked with every other vertex also called the complete graph, adjacency matrices are highly sought after Probing various Netting attributes could yield mighty insight into a biological Netting’s internal organization of molecules that are repartitioned between cellular processes or metabolic Nettings into a structure that is feasible and functional. For instance, consider a brief explanation of the fundamental attributes analyzed in nettings. The density of such structure reveals how it is as per the number of links per-vertex set and is set as = 2|E|/|V|(|V|–1). By a graph structure that is sparse we mean the one where O(|V|^k)=|E| and 1 < k<2 or |E| ∼ |V|. We also consider it dense when |E| ∼ |V|². It is known that nettings that are biological are frequently linked in a sparse way. For instance, in the regulatory Nettings that are transcriptional of S. cerevisiae, E. coli has a density less than 0.1 [37]. The complete graph structure has n(n–1) /2 arcs, with n=|V|. It is deemed regular with degree count |V| -1.

Let G₁ =(V₁, E₁) and G₂=(V₂, E₂). A function g: V₁ ⟶ V₂ is called an isomorphism if g is a bijection that preserves edges. That is for all r, s ∈ V₁, (r, s) ∈ E₁ if and only if (g(r), g(s)) ∈ E₂. A walk is a traverse along a given sequence of vertices (r₁,r₂,..., r_t) such that {(r₁,r₂),(r₂,r₃),..., (r_t - 1,r_L)} ⊆ E. A walk is called a simple path if no vertex appears more than once. A cycle is a closed path (r₁, r₂,...,r_t) where r₁ = r_t and t > 3. A path is called a trail if no edge appears more than once. A graph is called cyclic if a cycle is contained in it. Otherwise, it is named acyclic. An graph with no arc direction is considered to be connected any two elements are are separated by a path length count≥1. A di graph is deemed as strongly connected if it has directions marked for its arcs. The distance δ(r, s) from r to s is the path from r to s in G with least path length count. Else we set δ(r, s)=∞. For all practical purposes, we let δ(r, s)=∞=(max δ(r, s)+1). By the least path problem, we mean a procedure to find a path among two vertices so that the algebraic sum of edge weights is smallest. The mean path length and the diameter are considered as the expected and greatest δ(r, s) considered over all node pairs r, s ∈ V(G) with r≠s and are joined by at least one path. It is determined as δ=(2/ n(n–1)) varSigmavarSigma δ_min(r, s) where r, s = 1 to n and δ_min(r, s) is the least distance among r and s. The diameter of a netting is the longest least path within a graph and is set as D = max δ_min(r, s). Dijkstra’s greedy algorithm is a popular algorithm for finding the path of least length [16] and Floyd’s dynamic algorithm in [19] one another used more often. The former has O(n²) time complicity and gives the least length path count among a source node r and all other nodes in the netting and the latter has a O(n³) time complicity and requires an all-against-all array that consists of the distances of each node to every other node in the netting.

A clique is a complete subgraph G₁ of G. If a subset of any node set of G has no pair of adjacent nodes in it then we call it an independent set. The number of nodes in a clique is called its size. If a clique cannot be made further big by adding one another adjacent vertex then we call it a maximal clique. The task of finding a clique with maximum size in a given G belongs to the class NP-complete. So it is hard to find a step by procedure to find it [5]. A knowhow of these has a number of applications that are biological. That is: determining a) Genes co-expressed as groups of a given microarray in a dataset, b) matching 3-D structures of molecules [63, 65] etc., Several useful tools such as Clique Finder [42], MIClique [65] and Bioconductor [21] are noteworthy softwares for clique analysis. The tendency of a graph to get divided into assemblages are quantified by a measure called the Clustering Coefficient. A cluster is a subset of nodes with several arcs joining these nodes to each other. Suppose that r is a node with degree t in G that has no direction for its arcs and that there are m arcs among the t neighbors of r in G, then by the local assemblage coefficient of r in G we mean C_r = 2 m/ t(t–1). Note that 0≤C_r ≤ 1. The average assemblage coefficient of the whole netting C_average is given by C_average = (1/ n) varSigma E_i / t_i(t_r–1) where r = 1 to n and n=|V| is the cardinality or the number of vertices. A Netting is more likely to form clusters if C_r is close to 1. Observe that, a clique would emanate with C_r = 1. Nettings that are biological in general possess high C_r compared to random nettings, which demonstrates its nature that is modular. That is several processes that are cellular are explained by sub-collection of biomolecules that make an relation module. Since processes that are cellular are interlinked, the modules also become linked. But, as the linking molecules are very few the corresponding module overlap is also very low [2, 49].

In graph-based Netting theory, centralization is a measure that describes whether a graph has a star-like topology If its value is close to 1, then there is a fair chance for the Netting to look like a star-like topology. If its value is close to 0, then there is a fair chance for the vertices of the graph to have on average the same connectivity. For instance, consider a cycle on four vertices. Here every vertex is joined with two neighbors. It is determined as central measure=(n/n-2)((max (t)/(n-1)) –density) and central measure ∼ (max(t)/(n-1)) –density.

1.1 Motivating Factor 1

A familiar gene associations disorder that relates a Netting of disorders and disease genes provide a hunting ground to investigate in one unified graph-theoretic set up all familiar phenotype and disease-gene associations, pointing to the common genetic origin of different diseases. Proteins are highly expressed in various tissue which is encoded by genes. The disease-causing genes are part of the whole human interactome and exhibit a larger chance of physical association with the expressed products and transcripts. This finding advocates the presence of different cancer-related functionalities in the molecular Netting. However, the wholesome of disease-causing genes are insignificant and exhibit less inclination towards protein translation. Also, the expression pattern of genes states that the genes are localized in the periphery of the functional Netting. Authors noted variations in gene expression of essential and disease-causing genes and such genes have undergone somatic mutations that cannot be peripheral in the Netting [26].

1.2 Motivating Factor 2

In [9] the authors reported a new technique depending on the graph evolution process to find a replica of glioma considered as the malignancy of brain cancer. They probed the phases that a graph walks through in its evolution and exhibited an imminent link between the graph phases and the malignancy of cancer. They built cell graphs from the photomicrographs of tissues, which are categorized as cancerous -normal, low-grade, and high-grade depending on the position of cells. They used the random phenomenon to create an edge between every pair of cells based on the Euclidean distance among them. They extracted for a cell-graph, its link counsel with the attributes of its connected components to knowhow the phase of the cell-graph. They proved by toiling with samples of brain tissue extracted surgically from 12 affected cases that, cell-graphs created for several tissue types grow distinctly and that they showed different phase attributes to distinguish a tissue type from one another [9].

2 Efficiency measures on graphs

An efficiency measure among vertices in a graph was introduced in [36]. The efficiency among two vertices r and s is defined to be ɛ(r, s) = 1/d(r, s) for all i≠j. The global efficiency of a graph is defined as ɛ_g(G) = (1/ n(n–1)) _varSigmai ≠ _j ɛ(r, s). Observe that 0≤ɛ_g(G)≤1 with equality to 0 only when E(G) = ø and with equality to 1 when |E(G)| = n(n-1)/2. The local efficiency is defined as e_l = (1/n) _{varSigmai ∈G}e_g(G_i) where G_i is the subgraph induced by the neighbors of i.

For example, Consider P₅ the path on five vertices. Let P₅ = r(r,s)s(s,t)t(t,u)u(u,v)v. Then the distance among every pair of vertices is shown in Fig. 1.

Table 1
Calculation of distance among all pair of vertices for P₅

r s t u v

r 0 1 2 3 4

s 1 0 1 2 3

t 1 1 0 1 2

U 3 2 1 0 1

V 4 3 2 1 0

	r	s	t	u	v
r	0	1	2	3	4
s	1	0	1	2	3
t	1	1	0	1	2
U	3	2	1	0	1
V	4	3	2	1	0

Fig. 1

Block diagram of cancer.

Its average distance among vertices is avgdist (P₅) = (1/n(n-1)) varSigma δ(p, q) where p,q varies over the set {r, s, t, u, v}. So avgdist(P₅) = (2/5×4)) [4×1 + 3×2 + 2×3 + 1×4] = 2. Now ɛ(p,q) = 1/δ(p, q) for all p≠q implies the efficiency among the vertices I shown in Fig. 2.

Table 2

Calculation of Efficiency among all pair of vertices for P₅

	r	s	t	u	v
r	0	1	1/2	1/3	1/4
s	1	0	1	1/2	1/3
t	1/2	1	0	1	1/2
u	1/3	1/2	1	0	1
v	1/4	1/3	1/2	1	0

Fig. 2

Graphical representation of FALCON - AART.

The global efficiency ɛ_g(P₅) = (2/(5×4)) [(1/4)×1+(1/3)×2+(1/2)×3 + 1×4] = 77/120∼0.641.

Even after the advent of immunotherapy, we observe a lot of cases afflicted with cutaneous melanoma, a severe cancer type with a high mortality rate [27, 30].

3 Protein-protein interaction nettings

Proteins are vital macromolecules of cells that orchestrate signaling transduction and biological functions of living organisms. Two or more proteins interact with each other at the molecular level within a cell through hydrogen bonding, Van der Waals forces, electrostatic forces, and hydrophobic effect regulated by a series of biochemical events. Protein interaction Nettings help to uncover the role of characterized and uncharacterized proteins in protein complexes and cellular signaling pathways. Gaining in-depth knowledge in protein interaction is crucial as they play a significant role in controlling many molecular processes by creating genetic, epigenetic, and metabolic Nettings in any disease development at its molecular level.

Abnormal functions of proteins caused by genetic alterations or epigenetic modifications happened in the central dogma of life - transcription, and translation at the gene level can give rise to complex diseases, particularly cancer. The deregulated expression of proteins such as loss of expression or overexpression of proteins can transform a normal healthy cell into uncontrollably proliferating cancer cells. Any dysregulated process including DNA damage, unusual expression of proteins in cancer development can be unraveled by applying Netting biology in protein–gene interaction, protein-protein interaction studies. With the numerous existing data on protein interaction, Netting applications aids in understanding the molecular mechanisms of cancer. In the mathematical graph sense, create an abstract Netting representing every protein as nodes/vertices and the association between them edges. This mode of comprehending the molecular level of cancer development with the representation of the abstract Netting eases the understanding of intricate biological processes.

It is significant to note every minute detail of proteins in the Netting creation. Every protein is distinct and its function is dependent on its structure, physical and chemical properties. During the Netting development, the inclusion of the 3D structure of proteins and protein interface is essential. Proteins in general associate with one another through binding sites. Each protein has a limited sum of binding sites and the binding sites can be very specific to a particular protein or open to many proteins for interactions. Protein-protein interaction does not have defined binding sites like enzyme-substrate interactions. In a geometrical outlook, protein interface topographies are important in determining the specificity, strength, and affinity of protein-protein interaction. Having said that, the interaction Netting is categorized into single-interface and multi-interface to discriminate distinct and overlapping binding sites. The recent research probes the specificity, affinity, up-regulation, and down-regulation of cancer-causing proteins in the protein interaction Netting. For a better understanding of the development of cancer incorporate and integrate protein interface characteristics and protein expression profiles to investigate and compare the normal protein-protein interaction Netting and cancer-related protein-protein interaction Netting.

Many approaches of computational intent such as text-mining, gene fusion, phylogenetic profiling, conserved neighborhood, machine learning techniques, etc were created in a guess forprotein interaction Nettings to determine the distinct and overlapping gene sequences, the association of closely related proteins, presence and absence of proteins within the genome, activation, and non-activation of proteins, co-expression and co-localization within and across species. Even though several customary approaches aim atprobing proteins/genes, a close study of structures that are differential among cancer and normal PPI could be a viable notion to elicit a crucial message for unearthing the mechanisms of cancer and other dreaded cancer forms. It is be remembered that the accumulated details prove vital for diagnosing disease and mode of treatments.

Lately, attempts were made for making use of Netting analysis of protein interaction Nettings of cancers. For instance, in [44] it is explained that the cardinality of pivots in cancerous protein interaction netting could be brought down significantly in contrast with protein interaction netting in tissues that are normal. The authors in [67] probed the gene expression and predicted protein interactions employing 22 dissimilar graphs which are examined for factors for twenty-nine cancers. They have taken into account the magnitude of how cancerous netting differ from random nettings and observed the vast spread of pivot proteins remains unchanged in the cancer appearance. Comprehending the differences between normal and cancer tissues is not an easy task in systems biology. Graph related parameters of protein interaction were compared and analyzed for five different cancers such as bone, breast, colon, kidney, and liver and compared with its normal tissues. The results showed the protein interaction Nettings is much dense and contains more edges in cancer tissues compared to normal tissues. The cancer-causing proteins inclined towards a higher degree mean, higher betweenness and centrality, weaker clustering coefficient, and shorter shortest path length. These graphical parameters remarkably aided in differentiating cancer protein-protein interactions and normal protein-protein interactions. Further these parameters demonstrates deviances from those of their analogous random Nettings. To have an in-depth understanding of these parameters in the association patterns, the multilayer framework exposes sixty-three common proteins among all the disease data sets. Increased clustering coefficients are observed in these sixty-three proteins when compared to the whole Nettings [56, 64]. With the increased experimental protein interaction data, graph theory applications act as a pertinent tool in computational biology. The PPI with one another within an organism and across species are studied by employing large-scale and high-throughput methods. The public data sets of protein interactions that are presently available are obtained from the above-mentioned methods. The micro-RNAs are post-transcriptional regulators that control the expression of sensor proteins. These sensor proteins assumes a key part in the activation of cells that are cancer afflicted. The micro-RNAs regulate the proteins that are sensor based and their participation is evident in tumor penetration, metastasis, and cancer progression. From result analyses and knowing the cancer development at the fundamental level, advanced statistics and graphical techniques are utilized in concordance with the extensive data and literature survey on the web [57].

4 Ovarian cancer

Ovarian cancer is one of the reproductive malaises that is fatal. Cancer originates and progresses in the pithelial cells ovaries, germ cells, and the sex cord-stromal cells, and epitheilial cancer is the common pathological type [22, 46]. Ovarian cancer is responsiblefor 4% of all cancers among women and ranks 4th as a deadly disorder [66]. This is because the disease is not diagnosed early in about 70% of women with the epithelial until it reaches Stage III of spreading to the upper abdomen. So it is pertinent to detect it early. Also, the treatment processrelies on the type and the stage to which it has progressed [1, 29]. CA- 125 also called Cancer-Antigen 125 test [54], is a blood test that finds the antigen level in the blood, and is widely recognized as a tumor marker. 80–90% of women with the disease will be in their later stages and hence will portray signs of blood with antigen content. The survival is higher when ovarian cancer is detected at the localized stage or before cancer has spread to nearby lymph nodes and other organs.

Ovarian cancer is caused by multiple factors incuding genetics, environmental, and lifestyle changes. It includes subtypes with multiple roots of origin, various risk factors, genetic mutations, biological behaviors, and prognoses. Much more data remains to be unraveledonthe Internet of Medical Things-IoMT. Finally, less attention has been set on the research work that concentrated to increase therapeutic intermediations by subtype or on the best means to lower the morbidity of ovarian cancer. It is surprising to observe that not much care is setto comprehending survivorship concerns and the regular care needs of women with ovarian cancer. This also includes the employment of IoMT of the physical symptoms, treatment, and psychosocial impacts of confirmation and treatment. Major subtypes of ovarian carcinoma are named depending on how the tumor cells take over the healthy cells that surface the various organs in the female genitourinary tract in IoMT. It is advocated that a proper classification system is set to be in the mode because of tumor tissue genes.

Presently, Cancer is a major research area in the therapeutic field. Correct predictions of non-similar tumor types possess greater value allowing for better treatment and lowering the risk of destructiveness in patients. Medical data were collected from health informatics repositories wherein a massive volume of data is stored up. This Ovarian cancer database is utilized for the detection and process- identification. The dataset consists of non similar amounts of characteristics and owns missing data and noise for better ovarian cancer detection. Some features were selected with the help of a self-organizing map via an unsupervised learning process to determine a collection of structures in a dataset. The data set comprise huge data that is hard to pre-process. Hence the feature subset must be chosen in advance to handle the features. From the picked features to the supervised learning approach, it was accepted that recurrent neural Netting-RNN classifier was handy to recognize ovarian cancer disease as malignant or benign. The model should be modified for performance so that proper optimization is done to optimize the weights in RNN format and data are classified. Certainviable data could lead to classification procedure and will be the reason for settling the cancerous tissues. Such a system is graphically portrayed in Fig. 1.

5 Fuzzy neural netting-FNN for ovarian cancer

FNN borrows concepts and notions from fuzzy logic and neural netting. So, it can model the problem space far better and give human-like reasoning for its output. FNN is successfully applied in several areas as a decision support system. Fuzzy Adaptive Learning Control Netting-FALCON gives raise to another Adaptive Resonance Theory called (FALCONAART). This Netting can derive and formulate without prior knowledge of the task domain a fuzzy rule-base automatically for ovarian cancer diagnosis. It independently generates fuzzy rules of the form If x₁ is A and x₂ is B; Then y₁ is C and y₂ is D. It is an instance of a system with two inputs and two yields. a) Input variables (x₁,x₂) that is linguistic, b) Input terms (A, B) that is linguistic. It stands for entities that are fuzzy such as short, tall, fat etc., FALCON-AART represents gain terms that is linguistic by membership function that is graphical to denote the these terms. The If-Then rule: links the gain variables that is linguistic and terms as above with the consequent part variables that is linguistic and terms as below). Then yield variables (y₁, y ₂ ) that is linguistic. Finally Yield terms (C, D) that is linguistic.

If FALCON-AART is employed as a Stage II ovarian cancer classifier, then Stage II (positive) samples and non-Stage II (negative) samples will be employed to train the system. So when Stage II samples are presented to the netting, positive (Stage II) rules will have superior matching than negative (non-Stage II) rules. As a result, positive rules are activated and at the same time, negative rules are inhibited, leading to a positive decision. Each netting is trained with a different set of testing/training data (1/3 of data for training, 2/3 of data for testing). Each class is trained using three random sets of training/testing data and then the classification result is averaged. It is depicted using a Graph as shown in Fig. 4.

FALCON-AART is tremendously useful for diagnosing ovarian cancer. Fast training, simple fuzzy rule generation, and greater accuracy make FALCON-AART much sought after. A prominent feature is its capability to make complementary fuzzy rules for its reasoning process. These rules help clinicians in making their diagnostic decision. It substantially lessens the occurrence of medication errors and avoids the time-consuming process of deriving knowledge from large datasets such as microarray by hand. These decision support systems can then complement each other in ovarian cancer diagnosis, and this is believed to enhance diagnostic accuracy. Here again, Graphical tools aid us to visualize the connectivity pattern.

6 Influence of social netting

Social Netting enables us comprehend the strength of social support/relations on cancer patients-right from prevention to survival. One can evaluate a number of factors that are social in speculating how cancer patients exhibit their behaviours and identities when afflicted with cancer, and how these are influencing their social environment. Main factors are social status, family, support groups, health care peoples including those in online Nettings. Links with these various peoples spell the opportunities, constraints and resources. To achieve cancer prevention, peer advise group has recommended for the liberal intake of fruits and vegetables to minimize risk of cancer. It used analysis of social netting to find cliques between groups. Peer educators were picked depending on high peer index values for measures such as betweenness centrality, degree centrality that reveals social prominence. Finally, the peer education within each of the cliques to exploit social reinforcement. Also see [3 , 33–35].

Online interactions even though provide a less personal set of interactions, but assumes a vital role in knowing how people take a decision regarding cancer treatment. Studies on a social Netting probes on Twitter and messages related to cancer, record how breast cancer was guessed with by in-degree centrality measure, the followers count, betweenness centrality, and closeness centrality etc., These reveals useful ideas regarding conceptualization of common opinion from the concerned people ars followed by other stakeholders of counsel may require revision to catch the types of interactions in absorbing, retelling and revising counsel and to comprehend the roles in dissemination over online that need not fit a structure of hierarchy. As impersonal links do not mark personally at a cancer patient, it is actually up to the affected to take decisions regarding how to entertain such interactions. Also see [4 , 36–67].

7 Conclusion

To what extent this signal transduction mechanism improve our apprehension and management of cancer? The same group of signalling proteins can showcase different reactions to external changes. The different responses regulate cell-fate decisions. Theoretical studies, graph combined computational modelling reveal novel characteristics of every small protein involved in signalling Nettings and drug/proteins in treatment therapeutics. Genome sequencing data gives an untold of genetics and epigenetics mutations in cancer. Different mutations may have the same or different functional accomplishments in molecular signalling. Literature surveys and theoretical studies help us to perceive cancer, how cells metastasize by avoiding cell death, regulate resistance to drugs, and recurrence after treatment. Alternatively, protein interaction Netting modelling discloses nodes of enriched Netting susceptibility to genetic/epigenetic changes or drugs and suggests ways to inhibit resistance before it occurs. To sum up, we have seen various graph parameters and the graph-based Netting approach helps us to classify various cancer types and how Netting-dependent strategies lead to the proper understanding of cancer and its much-dreaded variations.

8 Overall review of this article

This paper shows how the graph theory and neural Nettings act as vital tools for understanding different types of cancer in particular, the ovarian cancer. Various Netting-dependent ways help to learn Nettings of drugs, targets, and diseases to reposition it for other new targets. We are searching for a unified viewpoint of graph theory procedures for Netting analysis. The protein-protein interaction Nettings and the ovarian cancer are comprehensively discussed here. The main subject of the paper are the Fuzzy Adaptive Learning Control Netting-FALCON and the Adaptive Resonance Theory (FALCON-AART). Such Nettings can automatically derive and formulate, without prior knowledge of the task, a modus-ponens fuzzy rule-base for ovarian cancer diagnosis. FALCON-AART is used as a Stage II ovarian cancer classifier and we are considering its performance in diagnosing ovarian cancer as tremendous. Thus we addressed all the concern of the anonymous reviewers.

9 Contribution of each of the authors

Yegnanarayanan is the main motivator and suggested the topic and provided support both technically and for overall improvement. Krithicaa Narayanaa has taken all responsibility for technical writing regarding various types of cancer and ovarian cancer as it is her area of specialization. She also improved the quality with block diagram. Then M.Anitha has contributed in the graphical interpretation of ovarian cancer and in the collection of relevant reference material. Rujita Ciurea helped with technical input and Luigi Geo Marceanu valuable suggestions for the over all writeup and presentation.

Footnotes

Acknowledgment

The first author gratefully acknowledge the National Board of Higher Mathematics, Department of Atomic Energy, Government of India, Mumbai for financial support by their grant no. 02011/10/21NBHM-(R.P)/R&D-II/8007/Date:13-07-2021

References

Anastasiadis

P.G.

, Romanidis

K.N.

, Polichronidis

, Koutlaki

N.G.

, Tamiolakis

and Simopoulos

, The contribution of rapid intraoperative cytology to the improvement of ovarian cancer staging, Gynecologic Oncology 86 (2002), 244–249.

Barabási

A.-L.

, Gulbahce

and Loscalzo

, Netting medicine: a Netting-based approach to human disease, Nature Reviews Genetics 12 (2011), 56–68.

Barnes

and Prior

, Examining the idea of ‘subversion’ in public services. In: Barnes M, Prior D, editors. Subversive Citizens: Power, Agency and Resistance in Public Services. Bristol: The Policy Press (2009). p. 3–14.

Bressler

and Popp

, Delineation of opinions on BRCA testing amongst orthodox Jewish thought leaders, Nurs Res (2016), 65:E92.

Bron

and Kerbosch

, Algorithm 457: finding all cliques of an undirected graph, Commun ACM (ACM) 16(9) (1973), 575–577.

Barnes

, Morrill

, Taren

, Aickin

, Sennott-Miller

, Duller

, et al. Randomized trial testing the effect of peer education at increasing fruit and vegetable intake, J Natl Cancer Inst. 91 (1999), 1491–500. doi: 10.1093/jnci/91.17.1491

Burgos

, Ceva

, Hernández

, Perazzo

R.P.J.

, Devoto

and Medan

, Two classes of bipartite Nettings: nested biological and social systems, Phys Rev (2008), 78.

Canty

, Cervical cancer screening among black women: a meta-ethnography, Nurs Res. 65 (2016), E23.

CigdemGunduz-Demir , , Mathematical modeling of the malignancy of cancer using graph evolution, Mathematical Biosciences, 209(2) (2007), 514–527.

10.

Cobb Sweeney

and Etkins

, Caregiver reactions in oncology and other chronic illnesses, Nurs Res. 65 (2016), E104.

11.

Cole-Lewis

, Perotte

, Galica

, Dreyer

, Griffith

, Schwarz

, et al., Social Netting behavior and engagement within a smoking cessation Facebook page, J Med Internet Res. 18 (2016), e205. doi: 10.2196/jmir.5574

12.

Cormen

T.H.

, Leiserson

C.E.

, Rivest Ronald

and Stein

, Introduction to algorithms. Cambridge, Massachusetts 02142: The MIT Press; 2002.

13.

Crookes

, Crookes

, Shelton

, Tehranifar

, Aycinena

, Gaffney

, et al. Social Nettings and social support for healthy eating among Latina breast cancer survivors: implications for social and behavioral interventions, J Cancer Surviv. 10 (2016), 291–301. doi: 10.1007/s11764-015-0475-6

14.

Cutrona

, Wagner

, Roblin

, Gaglio

, Williams

, Torres-Stone

, et al. E-mail to promote colorectal cancer screening within social Nettings: acceptability and content, J Health Commun. 20 (2015), 589–98. doi: 10.1080/10810730.2015.1012238

15.

Dean

L.T.

, Subramanian

S.V.

, Williams

D.R.

, Armstrong

, Charles

C.Z.

and Kawachi

, Getting black men to undergo prostate cancer screening: the role of social capital, Am J Mens Health 9 (2015), 385–96. doi: 10.1177/1557988314546491

16.

Dijkstra

E.W.

, A note on two problems in connexion with graphs, Numerische Mathematik, 1 (1959), 269–271.

17.

Durant

, McCray

and Safran

, Social Netting analysis of an online melanoma discussion group, Summit Transl Bioinform. 2010 (2010), 6–10.

18.

Ersig

, Hadley

and Koehly

, Understanding patterns of health communication in families at risk for hereditary nonpolyposis colorectal cancer: examining the effect of conclusive versus indeterminate genetic test results, Health Commun. 26 (2011), 587–94. doi: 10.1080/10410236.2011.558338

19.

Floyd

R.W.

, Algorithm 97, Comm ACM 5-6 (1962), 345.

20.

Freedman

, Viswanath

, Vaz-Luis

and Keating

, Learning from social media: utilizing advanced data extraction techniques to understand barriers to breast cancer treatment, Breast Cancer Res Treat. 158 (2016), 395–405. doi: 10.1007/s10549-016-3872-2

21.

Gentleman

R.C.

, Carey

V.J.

, Bates

D.M.

, Bolstad

, Dettling

, Dudoit

, Ellis

, Gautier

, Ge

, Gentry

, et al: Bioconductor: open software development for computational biology and bioinformatics, Genome Biology 5(10) (2004), R80.

22.

Gershenson

D.M.

, Management of early ovarian cancer: germ cell and sex cord-stromal tumors, Gynecologic Oncology 55 (1994), S62–S72.

23.

Ginex

, Ross

, Lugar

, Ross

, Lugar

, Tan

K.S.

, et al. From prescription to action: a mixed methodology approach to identify facilitators and barriers to adherence to oral cancer medications, Nurs Res. 65 (2016), E110.

24.

Hoybye

, Johansen

and Tjornhoj-Thomsen

, Online interaction. Effects of storytelling in an internet breast cancer support group, Psychooncology 14 (2005), 211–20. doi: 10.1002/pon.837

25.

Huber

, Carey

V.J.

, Long

, Falcon

and Gentleman

, Graphs in molecular biology, BMC Bioinformatics 8(Suppl 6) (2007), S8.

26.

Il Goh

, Cusick

M.E.

, Valle

, Childs

, Vidal

and Barabasi

A.L.

, The human disease Netting, Proc. Natl. Acad. Sci. U.S.A. 104(21) (2007), 8685–8690.

27.

, Yousaf

and Larkin

, Melanoma epidemiology, biology and prognosis. EJC Suppl EJC Off J EORTC, Eur Organ Res Treat Cancer. [et al], Elsevier 11 (2013), 81–91.

28.

Jensen

L.J.

, Kuhn

, Stark

, Chaffron

, Creevey

, Muller

, Doerks

, Julien

, Roth

, Simonovic

, et al: STRING 8–a global view on proteins and their functional interactions in 630 organisms, Database, Nucleic Acids Res 37 (2009), D412–416.

29.

Jermal

, Murray

, Samuels

, Ghafoor

, Ward

and Thun

M.J.

, Cancer statistic, CA Cancer J. Clin 13 (2003), 5–26.

30.

Karlsson

A.K.

and Saleh

S.N.

, Checkpoint inhibitors for malignant melanoma: a systematic review and meta-analysis, ClinCosmetInvestigDermatol. Dove Press 10 (2017), 325–39.

31.

Katz

, Pennell

, Dignan

and Paskett

, Assessment of cancer education seminars for Appalachian populations, J Cancer Educ. 27 (2012), 287–93. doi: 10.1007/s13187-011-0291-2

32.

Kholodenko

B.N.

, Hancock

J.F.

and Koch

, Signalling ballet in space and time, Nature Rev Molecular Cell Biology 11 (2010), 414–426.

33.

Kim

, Hou

, Han

J.Y.

and Himelboim

, Predicting retweeting behavior on breast cancer social Nettings: Netting and content characteristics, J Health Commun. 21 (2016), 479–86. doi: 10.1080/10810730.2015.110332

34.

Kreuter

, McBride

, Caburnay

, Poor

, Thompson

, Alcaraz

, et al. What can health communication science offer for ACA implementation? Five evidence-informed strategies for expanding medicaid enrollment, Milbank Q. 92 (2014), 40–62. doi: 10.1111/1468-0009.12040

35.

Kroenke

C.H.

, Quesenberry

, Kwan

M.L.

, Sweeney

, Castillo

and Caan

B.J.

, Social Nettings, social support, and burden in relationships, and mortality after breast cancer diagnosis in the life after breast cancer epidemiology (LACE) study, Breast Cancer Res Treat 137 (2013), 261–71. doi: 10.1007/s10549-012-2253-8

36.

Latora

and Marchiori

, Efficient behavior of small-worldnettings, Physical Review Letters 87(19) (2001).

37.

Leclerc

R.D.

, Survival of the sparsest: robust gene Nettings are parsimonious, MolSystBiol 4 (2008), 213.

38.

Lee

, McGlone

, Samg

, Avery

, Eubank

, Chang

, et al. Social Netting analysis of patient sharing among hospitals in Orange County, California. Am J Public Health 101 (2011), 707–13. doi: 10.2105/AJPH.2010.202754

39.

Lee

H.K.

, Hsu

A.K.

, Sajdak

, Qin

and Pavlidis

, Coexpression analysis of human genes across many microarray data sets, Genome Res 14(6) (2004), 1085–1094.

40.

Long

, Cunningham

, Wiley

, Carswell

and Braitwaite

, Leadership in complex Nettings: the importance of Netting position and strategic action in a translational cancer research Netting, Implement Sci. 8 (2013), 122. doi: 10.1186/1748-5908-8-122

41.

Long

, Cunningham

and Braithwaite

, Netting structure and the role of key players in a translational cancer research Netting: a study protocol, BMJ Open 2 (2012), e001434. doi: 10.1136/bmjopen-2012-001434

42.

Manfield

I.W.

, Jen

C.H.

, Pinney

J.W.

, Michalopoulos

, Bradford

J.R.

, Gilmartin

P.M.

and Westhead

D.R.

, Arabidopsis Co-expression Tool (ACT): web server tools for microarray-based gene expression analysis, Nucleic Acids Res 34 (2006), Web Server: W504–509.

43.

Marcus

, Understanding barriers to brca1/2 screening among Hasidic women in brooklyn, Nurs Res. 65 (2016), E31.

44.

Mirzarezaee

, Araabi

B.N.

and Sadeghi

, Comparison of hubs in effective normal and tumor protein interaction nettings, Basic and Clinical Neuroscience 2(1) (2010), 44–50.

45.

Picard

, Miele

, Daudin

J.-J.

, Cottret

and Robin

, Deciphering the connectivity structure of biological Nettings using MixNet, BMC Bioinformatics (2009), 10.

46.

Platzer

, Perco

, Lukas

and Mayer

, Characterization of protein-interaction Nettings in tumors, BMC Bioinformatics 8 (2007), 224.

47.

Polek

and Hardie

, Does having a family member diagnosed with cancer change fatalistic beliefs in Asian Americans? Nurs Res. 65 (2016), E1.

48.

Ramanadhan

, Salhi

, Baril

, D’Entermont

, Grullom

, Judge

, et al. Addressing cancer disparities via community Netting mobilization and intersectoral partnerships: a social Netting analysis, PLoS ONE 7 (2012), e32130. doi: 10.1371/journal.pone.0032130

49.

Ravasz

, Somera

, Mongru

, Oltvai

and Barabási

A.-.L.

, Hierarchical organization of modularity in metabolic Nettings, Science 297 (2002), 1551–1555.

50.

Schoenberg

, Eddens

, Jonas

, Snell-Rood

, Studts

, Broder-Oldach

, et al. Colorectal cancer prevention: perspectives of key players from social Nettings in a low-income rural US region, Int J Qual Stud Health Well-Being 11 (2016), 30396. doi: 10.3402/qhw.v11.30396

51.

Schulz

H.-J.

, John

, Unger

and Schumann

, Visual analysis of bipartite biological Nettings. Eurographics Workshop on Visual Computing for Biomedicine 2008.

52.

Selby

, van Mierlo

, Voci

, Parent

and Cunningham

, Online social and professional support for smokers trying to quit: an exploration of first time posts from members,, J Med Internet Res 12 (2010), e34. doi: 10.2196/jmir.1340

53.

Sinding

, Hudak

, Wiernikowski

, Aronson

, Miller

, Gould

, et al. I like to be an informed person but negotiating responsibility for treatment decisions in cancer care, Soc Sci Med. 71 (2010), 1094–101. doi: 10.1016/j.socscimed.2010.06.005

54.

Sliutz

, Templer

, Kainz

, Mustafa

, Gitsch

, Koelbl

and Beigelmayer

, Tissue polypeptide specific antigen and cancer-associated serum antigen in the follow-up of ovarian cancer.

55.

Sternas

, Women’s perceptions of their breast cancer experience, coping, health promotion activities and social support, Nurs. Res. 65 (2016), E98.

56.

Sun

and Zhao

, A comparative study of cancer proteins in the human protein-protein interaction Netting, BMC Genomics 11(Suppl 3) (2010), S5.

57.

Tabassum

D.P.

and Polyak

, Tumorigenesis: it takes a village, Nat. Rev. Cancer. 15 (2015), 473–483.

58.

Tsalatsanis

, Barnes

, Hozo

, Skvoretz

and Djubegovic

, A social Netting analysis of treatment discoveries in cancer, PLoS ONE 6 (2011), e18060. doi: 10.1371/journal.pone.0018060

59.

Valente

, Fujimoto

, Palmer

and Tanjasiri

, A Netting assessment of community-based participatory research: linking communities and universities to reduce cancer disparities, Am J Public Health 100 (2010), 1319–25. doi: 10.2105/AJPH.2009.171116

60.

Valente

T.W.

, Netting interventions, Science 337 (2012), 49–53. doi: 10.1126/science.1217330

61.

Van der Eerden

, Csikos

, Busa

, Hughes

, Radbruch

, Menten

, et al. Experiences of patients, family and professional caregivers with integrated palliative care in Europe: protocol for an international, multicenter, prospective, mixed method study, BMC Palliat Care 13 (2014), 52. doi: 10.1186/1472-684X-13-52

62.

Vanderveen

, Paterniti

, Kravitz

and Bold

, Diffusion of surgical techniques in early stage breast cancer: variables related to adoption and implementation of sentinel lymph node biopsy, Ann Surg Oncol. 14 (2007), 1662–9. doi: 10.1245/s10434-006-9336-x

63.

Voy

B.H.

, Scharff

J.A.

, Perkins

A.D.

, Saxton

A.M.

, Borate

, Chesler

E.J.

, Branstetter

L.K.

and Langston

M.A.

, Extracting gene nettings for low-dose radiation using graph theoretical algorithms, PLoSComputBiol 2(7) (2006).

64.

Xiong

, Xie

, Zhou

, Liu

and Guan

, The centrality of cancer proteins in human protein-protein interaction Netting: a revisit, International Journal of Computational Biology and Drug Design 7(2/3) (2014), 146–156.

65.

Zhang

, Song

, Wang

and Zhang

, MIClique: An Algorithm to Identify Differentially Coexpressed Disease Gene Subset from Microarray Data, Journal of Biomedicine and Biotechnology 2009.

66.

American Cancer Society. URL: http://www.cancer.org

67.

National Ovarian Cancer Coalition. URL: http://www.ovarian.org