Exploring Computational Methods to Identify Inhibitors for CDK2 Kinase

Exploring Computational Methods to Identify Inhibitors for CDK2 Kinase | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Exploring Computational Methods to Identify Inhibitors for CDK2 Kinase Rajneet Kaur Bijral, Inderpal Singh, Jatinder Manhas, Varun Sharma, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4315752/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 3 You are reading this latest preprint version Abstract Deregulation of Cyclin-Dependent kinase 2 is one of the main causes of lung cancer and is considered as a potential target for cancer therapy. Discovery of the inhibitors against the kinase using traditional methods is costly and time consuming. This issue can be easily dealt with by the usage of potential computational methods. Molecular Dynamic simulation has the ability to offer scientists an in-depth understanding of the dynamics of biomolecules at an atomic level. In the present study the molecular dynamic simulation of Cyclin Dependent kinase is carried out for one microsecond. T-sne and Isomap dimensionality reduction techniques are used followed by different clustering algorithms; Hierarchical, K-means, PAM, Hierarchical K-means and Clara. Subsequently, an optimal number of clusters was generated by using the validation index; Average Silhouette Width (ASW), Dunn Index and Davies Bouldin Index. Further, the cluster representative for each cluster is virtually screened with 52,000 drug-like molecules taken from May bridge library. Based on docking score the top 6 drugs like molecules are selected. Finally, the ADMET properties of the top compounds are evaluated using the Swiss ADMET tool. The strong binding energies and ADMET properties suggest that these molecules are considered to be valuable inhibitors and play a vital role in drug discovery. CDK2 kinase Isomap T-sne Clustering Simulation Figures Figure 1 Figure 2 1. Introduction CDK2 also known as Cyclin-dependent kinase 2 is the enzyme encoded in the CDK2 gene belonging to the family of Ser/Thr protein kinases. The enzyme plays an important role in the cell division of eukaryotic cells by catalyzing the transfer of the phosphoryl from the adenosine-5′-triphosphate (ATP) γ-phosphate to serine or threonine hydroxyl in substrate of the protein. The activation of CDK is the two-step process involving the cyclin binding and phosphorylation of threonine residue located in the kinase activation loop. Over-expression or deregulation of CDK2 causes the abnormal division of cells that is directly associated with the hyper-procreation of cancer cells that may cause various cancers; lung carcinoma, ovarian carcinoma, melanoma, osteosarcoma and pancreatic carcinoma [ 1 ][ 2 ]. Thus CDK2 is considered as a potential target for cancer therapy. The knowledge of CDK2 structure and availability of its X-ray crystal structure facilitates the identification of various CDK2 inhibitors that inhibit the signal pathway of the kinase, thus helping to cure cancers. The static model produced by the X-ray crystallography and NMR provide the information about the macromolecular structure of the kinase but drug binding is the dynamic process that is not attained well from these static structures alone. Kinase exhibits dynamic behavior and undergoes various conformational changes inside the cell and the drug binds stably with only few conformations. Thus considering the receptor motion and dynamics has a vital role in drug discovery. To understand the structural behavior and dynamics of the kinase at an atomic level Molecular Dynamic (MD) Simulation is used [ 3 ]. MD simulation captures and predicts the motion of atoms evolved over time depends on interatomic interactions governed by laws of physics [ 4 ]. MD simulations capture a variety of bio-molecular processes including; protein folding, ligand binding, response of bio-molecules to protonation, phosphorylation etc. It can generate thousands of conformations of protein structures while each conformation contains thousands of atoms. A lot of time and knowledge is required to analyze the large amount of data generated by MD simulations. The problem can be solved by applying the computational techniques [ 23 ] to analyze the trajectory data generated by MD [ 5 ]. Various dimensional reduction techniques and clustering techniques of Artificial Intelligence (AI) can be employed on trajectory data to reduce the dimension and to partition and classify the frames generated by simulation. In the present study dimensional reduction techniques T-sne and Isomap was applied on the trajectory of data generated by the MD simulation to decrease the dimension of the data. Then different clustering techniques: Hierarchical, K-means, PAM, Hierarchical K-means and Clara are applied by varying the number of clusters from 2 to 10. The indexes; Average Silhouette Width (ASW), Dunn Index and Davies Bouldin Index are used to find out the optimal number of clusters. Then for each cluster one representative is created to represent the clusters generated. 2. Methodology 2.1 Molecular Dynamic Simulations The 6lnl PDB id ( www.rcsb.org ) was taken from the protein data bank [ 6 ]. The hydrogen atoms missing in the structure, missing backbone atoms and residues were added by protein preparation wizard of Schrodinger Maestro software (Schrodinger Inc.). Energy minimization was done to correct the steric clashes in the 3D structure. The amino acids side-chain optimal protonation was achieved by setting pKa value of pH 7.2 by using Propka tool [ 7 ]. The structure was optimized again after protonation. The structure was positioned in the centre of the orthogonal box and then solvated with molecules of water using TIP3P water model [ 8 ]. 10 Angstroms was set as the periodic boundary condition and any surface atom of kinase was not at less than 10 Angstrom distance from the orthorhombic box boundaries. The prepared system had a charge and it was neutralized by addition of number of equal counter ions i.e. Na+/Cl- to neutralize the system. The physiological salt concentration was applied by addition of 150 millimolar NaCl. The system was energy minimized again after solvation. OPLS 2005 Force Field [ 9 ] was used for simulation. Constant pressure of 1 atm bar and temperature of 310 Kelvin were used using Berendsen Barostat to run the simulation. The time step of MD iteration was set to 2 femto-second using the SHAKE algorithm. To calculate the long range iteration Particle Mesh Ewald method was used. MD simulation of 1 microsecond time was calculated and recorded in the trajectory i.e., frame after every 1 nanosecond. The trajectory output was saved in .DTR and .CMS file format to store the trajectory and structure data respectively. 2.2 Dimension Reduction Techniques 2.2.1 T-sne (t-Distributed Stochastic Neighbor Embedding) T-sne (t-Distributed Stochastic Neighbor Embedding) [ 10 ] is the nonlinear dimensionality reduction technique that effectively projects the high dimensional data into low dimension and preserve the original data as much as possible. It used to embed high dimensional data into low dimensional data. T-sne is the variation of Stochastic Neighbor Embedding (sne) in which the similarity between the sample x i and x j for P-dimensional samples x 1 , x 2 , x 3 , .. x n is represented by finding out the conditional probability P i|j , where, i,j = 1, 2, 3, ...n. P i|j is high for the samples having high similarity but P i|j is almost zero for widely separated samples. The p j|i is given as, p j|i =exp(− d 2 (x i , x j )/2 σ 2 i )/∑ k 6≠i exp(− d 2 (x i , x k )/2 σ 2 i ) for i ≠ j, and p i|i = 0 , Where d 2 ( x i , x j ) represents the square of the Euclidean distance between the sample x i and sample x j and σ 2 i represents the variance of the Gaussian distribution, centred on sample x i . Probability in the original space is defined mathematically as; p j|i =( p j|i + p i|j )/ 2n for i ≠ j, and p ii = 0 , Where n is the size of the sample. In the t-SNE algorithm “perplexity” is an input parameter that specifies the number of neighbours. It can be mathematically defined as perp(P i ) = 2 H(Pi) Where H(Pi) is the Shannon entropy. T-sne employs student t- distribution (with single degree of freedom) to avoid overcrowding. 2.2.2 Isomap Isomap is the nonlinear dimensionality reduction technique that preserves the geodesic proximities using non-euclidean distances [ 11 ]. The geodesic metric precisely preserves the inter-point distances if the data lies in the nonlinear manifold. The isomap algorithm works in three steps. In the first step inter-point distance x ij is constructed for neighbouring points. This is done by considering K-nearest points for the target point. Then the weighted graph is used to represent the distance x ij , where weight of the edge defines the distance for all i and j. The second step is to define the geodesic distance g ij for two points by using shortest path in the graph. The dissimilarity matrix Δ =(g ij ) is constructed then the last step of isomap is done by applying multidimensional scaling(MDS) to find out the embedding space and associated principal components. 2.3 Clustering Clustering is the unsupervised learning technique to partition the data without any prior knowledge of the data and it provides more insight about the complex dataset. In clustering the identical groups or partitions are identified and grouped into the same group [ 12 ]. Different types of clustering algorithms are; Hierarchical clustering, Partition clustering and density based clustering. In hierarchical clustering the hierarchy of clusters is formed based on the similarity in the data. The hierarchy of data is presented in the form of a tree called dendrogram. Two approaches of hierarchical clustering are; agglomerative and divisive. In agglomerative every data point is considered as a single cluster and then the clusters are merged based on some similarity index. This process continues until one cluster is formed. The divisive algorithm is the reverse of the agglomerative approach. In partition clustering the data points are partitioned into a user-specified number of clusters (K) based on some similarities among data. Some of the popular partition clustering algorithms are; K-Means [ 13 ], PAM (K-Medoids)[ 14 ], CLARA (Clustering Large Applications)[ 15 ] etc. 2.4 Validation Index 2.4.1 Average Silhouette Width Average Silhouette Width (ASW) [ 16 ] is the internal validation index used to find out the optimal number of clusters. It is obtained by the difference between average distance between within the cluster and minimum distance between the clusters. The number of clusters for which ASW is maximum, is the optimal number of clusters. Mathematically calculated as: 𝑆 = 1/ (( ∑ n i=1 ( 𝑏(𝑖) − 𝑎(𝑖))/ (𝑚𝑎𝑥{𝑎(𝑖), 𝑏(𝑖)})) Where, a(i) represents the average distance within the cluster, b(i) represents the minimum distance between the clusters. The number of clusters for which ASW is maximum is the optimal number of clusters 2.4.2 Davies Bouldin Index The Davies–Bouldin(DB) index[ 17 ] is calculated as the ratio between “within-cluster” and “between-cluster” distances. The number of clusters for which DB index is minimum, is the optimal number of clusters. Mathematically calculated as: DB(C) = 1/k(∑ k i=1 max j≤k,j≠i C ij ), k=|c| Where, C ij is the "ratio of within-to-between cluster distance" for the i th and j th clusters. C ij is mathematically calculated as: C ij =(C’ i +C’ j )/C ij Where, C’ i and C’ j is the average distance between every point in clusters and their centroids. C ij is the distance (Euclidean distance) between the two cluster centroids. 2.4.3 Dunn Index Dunn Index[ 18 ] is used to identify the well separated, compact clusters and small variance clusters. Higher the value of the Dunn index the better is the clustering technique. It can be calculated as: Dunn index (U) = min 1≤i≤c {min 1≤j≤c, j≠i {(𝛿( C i , C j ))/ (max 1≤k≤c (𝛿(C k ))}} Where, ( C i , C j ) is the inter-cluster distance between C i and C j . 𝛿( C k ) is the intra-cluster distance within cluster C k . 3. Result and Discussion X-Ray Crystallographic structure of the protein provides details of its 3D architecture and information about its binding pocket, as such insights are static in nature, and the dynamic nature of the protein cannot be understood well through these methods alone. Besides, dynamic behaviour of the protein plays a vital role in the drug discovery process. It can be well understood by MD simulation. In this study, 1 microsecond long MD simulation of CDK2 Kinase done and 1000 conformations were generated using the Desmond MD tool. All the frames of the MD trajectory were aligned to null down the translations. The alpha carbon atoms of each of the amino acids were selected as the representative atom for the root mean square deviation (RMSD) analysis as shown in Fig. 1 . Through the RMSD plots it was observed that the mean deviation was ~ 3.5 angstroms, N lobe amino acids deviated more than the C lobe amino acids. This suggests that there was considerable deviation in the kinase structure during the course of the MD. This prompted us to view the contribution of each amino acid in the fluctuation during the MD using the root mean square fluctuation (RMSF) analysis. The RMSF of the Cdk2 is shown in Fig. 1 . The pattern of the local fluctuation profile of the amino acids was in sync with the already published literature [ 19 ]. The loops recorded more deviations; the N terminal residues of the C helix also had relatively higher fluctuation [ 20 ]. The most fluctuations were observed in amino acids mapped on activation loop and the GHI sub helical domain. The E and F helix had least fluctuations. Following the deviation and fluctuational analysis, T-sne and Isomap were applied as dimensional reduction techniques to reduce the dimensionality of data. Then different clustering techniques; hierarchical, K-Means, HK-Means, PAM and CLARA were applied to cluster the data. Different indexes (Average Silhouette Width, Davies Bouldin Index and Dunn Index) were used to find out the best clustering algorithm and optimal number of clusters. Values of different indexes were recorded for five clustering techniques by varying the number of clusters from 2 to 10. T-sne was used for dimensionality reduction and different clustering techniques are applied to cluster the data. The value of Average Silhouette Width (ASW), Davies Bouldin(DB) Index and Dunn Index for different clustering techniques were recorded as shown in Table 1 , Table 2 and Table 3 respectively. It was observed from Table 1 , that all the clustering techniques have high ASW value when 10 clusters were formed. Out of 5 clustering techniques Hierarchical K-Means clustering recorded the highest value of 0.512444 (for 10 clusters). Table 2 shows the value DB index for different clustering techniques; hierarchical, K-Means, PAM and ClARA clustering with DB value of 0.690332, 0.6987556, 0.7072587 and 0.700863 respectively, showed 9 as the optimal number of clusters except HK-Means with DB value of 0,696726 showed 10 as optimal number of clusters. Table 1 Show the values of Average Silhouette Width (ASW) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction. Hierarchical Clustering K-means Clustering PAM Clustering HK-means Clustering Clara Clustering 2 Clusters 0.3885 0.3818612 0.4014626 0.388876 0.389238 3 Clusters 0.3962 0.4463395 0.4149279 0.465571 0.435407 4 Clusters 0.4297 0.4595713 0.4469391 0.459571 0.447717 5 Clusters 0.4654 0.4878133 0.4785397 0.482815 0.434354 6 Clusters 0.4831 0.4781707 0.4763377 0.491696 0.465199 7 Clusters 0.4861 0.4629503 0.4937016 0.495461 0.478229 8 Clusters 0.4848 0.476702 0.4758134 0.487784 0.465917 9 Clusters 0.4958 0.4958097 0.490095 0.49581 0.481096 10 Clusters 0.5005 0.4997089 0.5051389 0.512444 0.50783 Table 2 Shows the values of Davies Bouldin index (DB) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction. Hierarchical Clustering K-means Clustering PAM Clustering HK-Means Clustering Clara Clustering 2 Clusters 1.1258234 1.2367192 1.2049045 1.19084 1.222572 3 Clusters 0.8347396 0.8284007 0.9258379 0.806557 0.848357 4 Clusters 0.8051623 0.725148 0.7660555 0.725148 0.721648 5 Clusters 0.7605432 0.7083599 0.7115981 0.702764 0.743582 6 Clusters 0.7517587 0.7154494 0.7234466 0.708506 0.741091 7 Clusters 0.7643028 0.7899481 0.7079546 0.726907 0.74704 8 Clusters 0.7707523 0.722547 0.7272507 0.725646 0.732795 9 Clusters 0.690332 0.6987556 0.7072587 0.698756 0.700863 10 Clusters 0.7437819 0.7080912 0.7155078 0.696726 0.700888 Table 3 Show the values of Dunn Index (DI) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction Hierarchical Clustering K-means Clustering PAM Clustering HK-means Clustering Clara Clustering 2 Clusters 0.0243 0.0096 0.0101 0.00935 0.0036 3 Clusters 0.0311 0.0087 0.0112 0.008696 0.0072 4 Clusters 0.032 0.0057 0.0066 0.005683 0.0046 5 Clusters 0.011 0.0216 0.0092 0.006271 0.0026 6 Clusters 0.0136 0.0153 0.0061 0.015328 0.0099 7 Clusters 0.0136 0.0113 0.0077 0.01126 0.0112 8 Clusters 0.0136 0.016 0.0113 0.01126 0.0083 9 Clusters 0.0136 0.0184 0.0113 0.008241 0.0034 10 Clusters 0.0188 0.0184 0.0089 0.010273 0.0091 Value of the Dunn index is shown in Table 3 . Hierarchical, K-Means, PAM and ClARA clustering reported 0.032, 0.0216, 0.0113, 0.015328 and 0.0112 respectively as optimal value of Dunn index for 4, 5, 9, 6 and 7 respectively as the optimal number of clusters. When Isomap was used for dimensionality reduction and different clustering techniques are applied to cluster the data. The value of ASW for different clustering techniques was shown in Table 4 . It was observed that all the clustering techniques have high ASW value when 5 clusters are formed. Out of five clustering techniques; K-Means and Hierarchical K-Means clustering have recorded the highest value of 0.604545 (for 5 clusters). Table 5 shows that all the clustering techniques; hierarchical, K-Means, PAM, Hierarchical K-Means and CLARA clustering with DB value of 0.5996, 0.6045447, 0.6040297, 0.604545 and 0.599434 resp. show 5 as the optimal number of clusters. Table 6 shows that the value of DI for hierarchical clustering, K-Means, PAM Hierarchical K-Means and CLARA clustering is 0.047, 0.0571, 0.0236, 0.046064 and 0.0089 resp. for 10,7,8,10 and 3 resp. as the optimal number of clusters. It was observed that ASW, DB and DI had shown better values for different clustering techniques when Isomap was used as dimensional reduction techniques so for finding the optimal numbers of clusters Isomap was considered for dimensional reductionality. It was observed that K-Means clustering shows better index values (ASW and DB) for five as an optimal number of clusters. So we have considered five as the optimal number of clusters. Table 4 Show the values of Average Silhouette Width (ASW) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction. Hierarchical Clustering K-means Clustering PAM Clustering HK-means Clustering Clara Clustering 2 Clusters 0.5996 0.5997813 0.5962576 0.599781 0.594198 3 Clusters 0.5286 0.5777035 0.5802975 0.5783 0.580945 4 Clusters 0.5325 0.5864056 0.5899616 0.586406 0.589118 5 Clusters 0.5919 0.6045447 0.6040297 0.604545 0.599434 6 Clusters 0.5521 0.5648679 0.5692951 0.564868 0.555593 7 Clusters 0.5324 0.5438552 0.5313613 0.554465 0.528342 8 Clusters 0.5329 0.5234979 0.5314133 0.550032 0.512257 9 Clusters 0.5184 0.5223692 0.5339954 0.528485 0.508167 10 Clusters 0.5049 0.5085656 0.5172938 0.495137 0.494295 Table 5 Shows the values of Davies Bouldin index (DB) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction. Hierarchical Clustering K-means Clustering PAM Clustering Hk-means Clustering Clara Clustering 2 Clusters 0.6324226 0.6306561 0.6293577 0.630656 0.627538 3 Clusters 0.6024295 0.5863843 0.5747294 0.58534 0.571271 4 Clusters 0.5526272 0.5764032 0.55947 0.576403 0.56181 5 Clusters 0.546069 0.5450825 0.5462048 0.545083 0.545521 6 Clusters 0.6005217 0.5959664 0.5877089 0.595966 0.605961 7 Clusters 0.5929847 0.6129836 0.6154011 0.603058 0.624783 8 Clusters 0.5956353 0.6551261 0.6121824 0.615468 0.645534 9 Clusters 0.6129884 0.639297 0.6228808 0.6345 0.642229 10 Clusters 0.624127 0.6810369 0.6215201 0.673057 0.637879 Table 6 Show the values of Dunn Index (DI) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction. Hierarchical Clustering K-means Clustering PAM Clustering HK-means Clustering Clara Clustering 2 Clusters 0.0177 0.0034 0.0183 0.003439 0.0006 3 Clusters 0.0168 0.0047 0.0174 0.016919 0.0089 4 Clusters 0.02 0.028 0.0077 0.025194 0.0034 5 Clusters 0.0291 0.0397 0.0079 0.039718 0.0029 6 Clusters 0.0375 0.0438 0.0071 0.040631 0.0058 7 Clusters 0.0376 0.0571 0.0071 0.008078 0.0068 8 Clusters 0.0422 0.0446 0.0236 0.044116 0.0058 9 Clusters 0.0447 0.045 0.007 0.040499 0.0038 10 Clusters 0.047 0.0567 0.0098 0.046064 0.0036 For docking five frame representatives were created by finding out the average structure generated by computing the average mean of the Cartesian coordinates of each amino acid in all the structures present in a cluster. Then this average structure of each cluster was further used to identify the representative for each cluster. And the Structure having minimum deviation with average structure was found and saved as frame representative for each cluster. The representations thus generated were used in ensemble docking. The frame representatives thus generated were docked with drug-like molecules taken from 52,000 compounds of Maybridge Library (( http://www.maybridge . com)) using molegro virtual docker software [ 21 ]. The ranking of the drug molecule was done on the basis of docking energies derived by the physico-chemical interactions, shape complementarity and entropy. The top 6 compounds with high docking scores were listed in Table 7 . All these compounds cleared the Lipinski’s filter and had molecular weight between 334.4 and 463. Their rotatable bonds were in the range of 3 to 5 and that were checked using the Swiss ADME tool [ 22 ] for the ADME properties that is shown in Fig. 2 . Table 7 The top 6 compounds with strong binding energies in terms of Molecular Docking Score was shown S. No ZINC compound ID rotatable Number of rotatable bonds MolDock Score Molecular weight (g/ mol) IUPAC name 1 ZINC12369719 3 -136.131 376.4 5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole 2 ZINC04394657 5 -135.565 394.9 N-[(2-chlorophenyl)methylideneamino]-4-(5-methyl-3-phenyl-1,2-oxazol-4-yl)-1,3-thiazol-2-amine 3 ZINC01047457 3 -135.456 398.3 2-(3,4-dichlorophenyl)- N '-(thiophene-2-carbonyl)-1,3-thiazole-4-carbohydrazide 4 ZINC00169253 4 -132.29 334.4 4-[2-(2,1,3-benzoxadiazol-5-yloxymethyl)-1,3-thiazol-4-yl]benzonitrile 5 ZINC13658326 5 -130.353 463 4-[(Z)-[3-(4-chlorophenyl)-1-phenylpyrazol-4-yl]methylideneamino]-3-thiophen-2-yl-1H-1,2,4-triazole-5-thione 6 ZINC12369719 3 -127.131 376.4 5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole 4. Conclusion In the present study, molecular dynamic simulation of CDK2 was done and its diverse 3D conformations were generated. T-sne and Isomap were used as dimensional reduction techniques and then different clustering algorithms including Hierarchical, K-means, PAM, Hierarchical K-means and Clara were applied. Then validation indexes including Average Silhouette Width, Dunn Index and Davies Bouldin Index were used to evaluate dimensional reduction and different clustering techniques thus applied. Then the optimal number of clusters is found out, to find out the representative structure required for docking. Then the representative structures are virtually screened with the Maybridge database. Six novel target molecules were identified with promising binding energies and their ADMET properties were evaluated. Top 6 selected compounds: ZINC12369719, ZINC4394657, ZINC1047457, ZINC0016925, ZINC13658326 and ZINC12369719. The compounds thus found have promising scope for wet lab screening and design of inhibitors against CDK2. Declarations Author contributions R.K, I.S, J.M, Va.S and Vi.S conceived the idea. R.K, I.S and Va. S executed the experiment and wrote the article. J.M and Vi.S did edition and corrections. Data availability Not applicable. Code availability Not applicable. Funding: No funding Declarations Conflict of interest The authors declare no competing interests. Author’s Biography Rajneet Kaur Bijral completed her Ph.D in Computer Science from Department of Computer Science and IT, University of Jammu. She has an experience of over 5 years of teaching experience. Her area of interest includes Artificial intelligence, machine learning, deep learning, Bioinfomatic etc. Dr. Inderpal Singh is a reputed researcher in the field of Bioinfomatics. His area of research includes cancer drug discovery, kinase structure, genetics of rare disorder, metagenomics and immuoinformatics. He completed his Ph.D in Biotechnology from SMVDU University and served as faculty in SKUAST before starting his research company Bioinfores where he teaches biological data science analysis to researcher and professionals across globe. His publications are published across most reputed international publication house. Dr. Varun Sharma is Head of Bioinformatics at NMC Genetics India Pvt. Ltd.(NMC Healthcare). He has done his Ph.D. from Shri Mata Vaishno Devi University and Postdoc from Birbal Sahni Institute of Paleosciences, Lucknow, GoI. His research work includes pharmacogenomics evaluation of different complex disorders. He has been awarded with young scientist fellowship for his work on Pharmacogenomics by Council of Lindau Nobel Laureates Germany. He is having more than 30 research articles in the peer reviewed journals and book chapters. With an experience of more than 5 years he has great hold on developing statistical models for different diseases in correlation with their phenotypes, Genome wide association analyses and functional mapping and annotation of genetic association. Dr. Jatinder Manhas completed his MCA and Ph.D. in Computer Science from Department of Computer Science and IT, University of Jammu. He is an eminent Researcher and his publications are published across most reputed international publication house. He has a vast experience of over 13 years in the field of networking, databases, and website design issues. Currently he is working as a Senior Assistant Professor in the Department of Computer Science and IT, Bhaderwah Campus of University of Jammu. His area of interest includes Artificial intelligence, machine learning, deep learning, medical image analysis, website design issues, etc. Dr. Vinod Sharma received his Ph.D. in Computer Science from Department of Computer Science and IT, University of Jammu. He is an eminent Researcher and his publications are published across most reputed international publication house. He has over 26 years of experience in teaching and research. Currently he is working as a Professor in the Department of Computer Science and IT, University of Jammu. Besides this, he is also appointed as Director, Ramnagar Campus, and University of Jammu. His area of interest includes Artificial intelligence, machine learning, deep learning, medical image analysis, medical diagnosis, etc. References S. Hu, A. V. Danilov, K. Godek, B. Orr, L. J. Tafe., J. Rodriguez-Canales, C. Behrens, B. Mino, C. A. Moran, V. A. Memoli, L. M. Mustachio, F. Galimberti, S. Ravi, A. DeCastro, Y. Lu, D. Sekula, A. S. Andrew, I. I. Wistuba, S. Freemantle, D. A. Compton, … Dmitrovsky, E. (2015). “CDK2 Inhibition Causes Anaphase Catastrophe in Lung Cancer through the Centrosomal Protein CP110,” Cancer research , vol 75 , no 10, pp. 2029–38. doi.org/10.1158/0008-5472.Can-14-1494 S. Akli, C. S. Van Pelt, T. Bui, L. Meijer, and K. Keyomarsi, "Cdk2 is required for breast cancer mediated by the low-molecular-weight isoform of cyclin E," (in eng), Cancer Res, vol. 71, no. 9, pp. 3377-86, May 1 2011, doi: 10.1158/0008-5472.Can-10-4086. J. D. Durrant and J. A. McCammon, "Molecular dynamics simulations and drug discovery," BMC Biology, vol. 9, no. 1, p. 71, 2011/10/28 2011, doi: 10.1186/1741-7007-9-71. A. Hospital, J. R. Goñi, M. Orozco, and J. L. Gelpí, "Molecular dynamics simulations: advances and applications," (in eng), Adv Appl Bioinform Chem, vol. 8, pp. 37-47, 2015, doi: 10.2147/aabc.S70333. G. Sliwoski, S. Kothiwale, J. Meiler, and E. W. Lowe, Jr., "Computational methods in drug discovery," (in eng), Pharmacol Rev, vol. 66, no. 1, pp. 334-95, 2014, doi: 10.1124/pr.112.007336. H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig, I. N. Shindyalov, P. E. Bourne, "The Protein Data Bank," (in eng), Nucleic Acids Res, vol. 28, no. 1, pp. 235-42, Jan 1 2000, doi: 10.1093/nar/28.1.235. M. Rostkowski, M. H. Olsson, C. R. Søndergaard, and J. H. Jensen, "Graphical analysis of pH-dependent properties of proteins predicted using PROPKA," (in eng), BMC Struct Biol, vol. 11, p. 6, Jan 26 2011, doi: 10.1186/1472-6807-11-6. M. Harrach and B. Drossel, "Structure and dynamics of TIP3P, TIP4P, and TIP5P water near smooth and atomistic walls of different hydroaffinity," The Journal of chemical physics, vol. 140, p. 174501, 05/07 2014, doi: 10.1063/1.4872239. D. Shivakumar, E. Harder, W. Damm, R. Friesner, and W. Sherman, "Improving the Prediction of Absolute Solvation Free Energies Using the Next Generation OPLS Force Field," Journal of Chemical Theory and Computation, vol. 8, pp. 2553–2558, 07/09 2012, doi: 10.1021/ct300203w. D. Kobak and P. Berens, "The art of using t-SNE for single-cell transcriptomics," Nature Communications, vol. 10, no. 1, p. 5416, 2019/11/28 2019, doi: 10.1038/s41467-019-13056-x. C. Wang, W. Fu, H. Huang, and J. Chen, "Isomap-Based Three-Dimensional Operational Modal Analysis," Scientific Programming, vol. 2020, p. 6348372, 2020/07/14 2020, doi: 10.1155/2020/6348372. C. Maione, D. R. Nelson, and R. M. Barbosa, "Research on social data by means of cluster analysis," Applied Computing and Informatics, vol. 15, no. 2, pp. 153-162, 2019/07/01/ 2019, doi: https://doi.org/10.1016/j.aci.2018.02.003 . S. Na, L. Xumin, and G. Yong, "Research on k-means Clustering Algorithm: An Improved k-means Clustering Algorithm," in 2010 Third International Symposium on Intelligent Information Technology and Security Informatics , 2-4 April 2010 2010, pp. 63-67, doi: 10.1109/IITSI.2010.74. D. Lei, Q.-s. Zhu, J. Chen, H. Lin, and P. Yang, "Automatic PAM Clustering Algorithm for Outlier Detection," Journal of Software, vol. 7, 04/27 2012, doi: 10.4304/jsw.7.5.1045-1051. T. Gupta and S. P. Panda, "Clustering Validation of CLARA and K-Means Using Silhouette & DUNN Measures on Iris Dataset," in 2019 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (COMITCon) , 14-16 Feb. 2019 2019, pp. 10-13, doi: 10.1109/COMITCon.2019.8862199. F. Batool and C. Hennig, "Clustering with the Average Silhouette Width," Computational Statistics & Data Analysis, vol. 158, p. 107190, 02/01 2021, doi: 10.1016/j.csda.2021.107190. J. Xiao, J. Lu, and X. Li, "Davies Bouldin Index based hierarchical initialization K-means," Intelligent Data Analysis, vol. 21, pp. 1327-1338, 11/15 2017, doi: 10.3233/IDA-163129. J. C. Bezdek and N. R. Pal, "Some new indexes of cluster validity," IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), vol. 28, no. 3, pp. 301-315, 1998, doi: 10.1109/3477.678624. T. Mohammad, S. Batra, R. Dahiya, MH. Baig, IA. Rather, JJ. Dong, I. Hassan, "Identification of High-Affinity Inhibitors of Cyclin-Dependent Kinase 2 Towards Anticancer Therapy," (in eng), Molecules, vol. 24, no. 24, Dec 15 2019, doi: 10.3390/molecules24244589. K. Huang, Y. H. Wang, A. Brown, and G. Sun, "Identification of N-terminal lobe motifs that determine the kinase activity of the catalytic domains and regulatory strategies of Src and Csk protein tyrosine kinases," (in eng), J Mol Biol, vol. 386, no. 4, pp. 1066-77, Mar 6 2009, doi: 10.1016/j.jmb.2009.01.012. G. Bitencourt-Ferreira and W. F. de Azevedo, Jr., "Molegro Virtual Docker for Docking," (in eng), Methods Mol Biol, vol. 2053, pp. 149-167, 2019, doi: 10.1007/978-1-4939-9752-7_10. A. Daina, O. Michielin, and V. Zoete, "SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules," Scientific Reports, vol. 7, no. 1, p. 42717, 2017/03/03 2017, doi: 10.1038/srep42717. D. P. Singh, G. Abhishek, and K. Baijnath, "DWUT-MLP: Classification of anticancer drug response using various feature selection and classification techniques," Chemometrics and Intelligent Laboratory Systems, vol. 225, p. 104562, 2022/06/15/ 2022, doi: https://doi.org/10.1016/j.chemolab.2022.104562. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editor assigned by journal 30 Apr, 2024 Submission checks completed at journal 29 Apr, 2024 First submitted to journal 24 Apr, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4315752","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":297026794,"identity":"762d6e89-641b-4f55-b894-0528527acc20","order_by":0,"name":"Rajneet Kaur Bijral","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA20lEQVRIie2PsQrCQAyG06VdDrumWPQVzkkHxVeJCOdSQRCcndrBiq/SR+ikSx0FS0HOpS4+gIKDbd3bcxO8b0j+IR9JADSa30QAMYCWtS4yH3yhmCwuFVRSzE9FKpuC0g/iXEr30jWd2yl6LBDsYEO1ipvQjBNb9vy2t0rD4jBMjlGtgkACiZFRKOLMCoXjvEGxZaWMfScR6UtJwc+WiY/WPlPbgrL8haY+88zM5ciaf7E9cX2GNNpZhzy9v4YdO9jWKwCMwAirwKvaMF5ixQCPKkiFaY1Go/lH3jRpQh/WEyjwAAAAAElFTkSuQmCC","orcid":"","institution":"University of Jammu","correspondingAuthor":true,"prefix":"","firstName":"Rajneet","middleName":"Kaur","lastName":"Bijral","suffix":""},{"id":297026798,"identity":"fdd3055c-30c3-4630-be6f-b67a5b5dccc8","order_by":1,"name":"Inderpal Singh","email":"","orcid":"","institution":"Bioinfores","correspondingAuthor":false,"prefix":"","firstName":"Inderpal","middleName":"","lastName":"Singh","suffix":""},{"id":297026802,"identity":"89604c6c-505a-4d66-ae00-841c47f3eadb","order_by":2,"name":"Jatinder Manhas","email":"","orcid":"","institution":"University of Jammu","correspondingAuthor":false,"prefix":"","firstName":"Jatinder","middleName":"","lastName":"Manhas","suffix":""},{"id":297026807,"identity":"13b28d9c-ad9d-43fa-8962-113df1c8806a","order_by":3,"name":"Varun Sharma","email":"","orcid":"","institution":"NMC Genetics India Pvt Ltd","correspondingAuthor":false,"prefix":"","firstName":"Varun","middleName":"","lastName":"Sharma","suffix":""},{"id":297026809,"identity":"a312c432-6ac0-47d3-a1a2-9bd0fbb46496","order_by":4,"name":"Vinod Sharma","email":"","orcid":"","institution":"University of Jammu","correspondingAuthor":false,"prefix":"","firstName":"Vinod","middleName":"","lastName":"Sharma","suffix":""}],"badges":[],"createdAt":"2024-04-24 05:57:54","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4315752/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4315752/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":55912199,"identity":"5071c8cc-81d3-45d7-9799-16b06d6c138c","added_by":"auto","created_at":"2024-05-06 08:12:21","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":144411,"visible":true,"origin":"","legend":"\u003cp\u003eThe heavy atom RMSD plot of CDK2 kinase(a). The kinase undergoes structural changes, and thus,maximum RMSD was observed to be ~3.5 A. The RMSF plot for CDK2 kinase (b). The dynamic cross-correlation motion analysis plot shows the amino acids with positive (blue) and negative (red) motion correlations (C).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4315752/v1/9a53c9d2bc93fb6c79a0191d.png"},{"id":55912200,"identity":"a410b160-d6d1-4daf-9ff7-b870dee2046f","added_by":"auto","created_at":"2024-05-06 08:12:21","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":121488,"visible":true,"origin":"","legend":"\u003cp\u003eThe rador plot showing the ADME properties of the top 6 selected compounds; ZINC12369719, ZINC4394657, ZINC1047457, ZINC0016925, ZINC13658326 and ZINC12369719 respectively. The rador plot shows the permissible values for six physiochemical prop\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4315752/v1/c6701c93c1b92b83b0c7e335.png"},{"id":55913113,"identity":"b7b3703d-2bea-4427-8e88-cdaf12e0da3a","added_by":"auto","created_at":"2024-05-06 08:20:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1653558,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4315752/v1/eaef8800-b46f-4c6a-b681-64dcc023f45f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Exploring Computational Methods to Identify Inhibitors for CDK2 Kinase","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eCDK2 also known as Cyclin-dependent kinase 2 is the enzyme encoded in the CDK2 gene belonging to the family of Ser/Thr protein kinases. The enzyme plays an important role in the cell division of eukaryotic cells by catalyzing the transfer of the phosphoryl from the adenosine-5\u0026prime;-triphosphate (ATP) γ-phosphate to serine or threonine hydroxyl in substrate of the protein. The activation of CDK is the two-step process involving the cyclin binding and phosphorylation of threonine residue located in the kinase activation loop. Over-expression or deregulation of CDK2 causes the abnormal division of cells that is directly associated with the hyper-procreation of cancer cells that may cause various cancers; lung carcinoma, ovarian carcinoma, melanoma, osteosarcoma and pancreatic carcinoma [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e][\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Thus CDK2 is considered as a potential target for cancer therapy. The knowledge of CDK2 structure and availability of its X-ray crystal structure facilitates the identification of various CDK2 inhibitors that inhibit the signal pathway of the kinase, thus helping to cure cancers. The static model produced by the X-ray crystallography and NMR provide the information about the macromolecular structure of the kinase but drug binding is the dynamic process that is not attained well from these static structures alone. Kinase exhibits dynamic behavior and undergoes various conformational changes inside the cell and the drug binds stably with only few conformations. Thus considering the receptor motion and dynamics has a vital role in drug discovery. To understand the structural behavior and dynamics of the kinase at an atomic level Molecular Dynamic (MD) Simulation is used [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. MD simulation captures and predicts the motion of atoms evolved over time depends on interatomic interactions governed by laws of physics [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. MD simulations capture a variety of bio-molecular processes including; protein folding, ligand binding, response of bio-molecules to protonation, phosphorylation etc. It can generate thousands of conformations of protein structures while each conformation contains thousands of atoms. A lot of time and knowledge is required to analyze the large amount of data generated by MD simulations. The problem can be solved by applying the computational techniques [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] to analyze the trajectory data generated by MD [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Various dimensional reduction techniques and clustering techniques of Artificial Intelligence (AI) can be employed on trajectory data to reduce the dimension and to partition and classify the frames generated by simulation.\u003c/p\u003e \u003cp\u003eIn the present study dimensional reduction techniques T-sne and Isomap was applied on the trajectory of data generated by the MD simulation to decrease the dimension of the data. Then different clustering techniques: Hierarchical, K-means, PAM, Hierarchical K-means and Clara are applied by varying the number of clusters from 2 to 10. The indexes; Average Silhouette Width (ASW), Dunn Index and Davies Bouldin Index are used to find out the optimal number of clusters. Then for each cluster one representative is created to represent the clusters generated.\u003c/p\u003e"},{"header":"2. Methodology","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 \u003cb\u003eMolecular Dynamic Simulations\u003c/b\u003e\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe 6lnl PDB id (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"http://www.rcsb.org\" target=\"_blank\"\u003ewww.rcsb.org\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.rcsb.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was taken from the protein data bank [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. The hydrogen atoms missing in the structure, missing backbone atoms and residues were added by protein preparation wizard of Schrodinger Maestro software (Schrodinger Inc.). Energy minimization was done to correct the steric clashes in the 3D structure. The amino acids side-chain optimal protonation was achieved by setting pKa value of pH 7.2 by using Propka tool [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. The structure was optimized again after protonation. The structure was positioned in the centre of the orthogonal box and then solvated with molecules of water using TIP3P water model [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. 10 Angstroms was set as the periodic boundary condition and any surface atom of kinase was not at less than 10 Angstrom distance from the orthorhombic box boundaries. The prepared system had a charge and it was neutralized by addition of number of equal counter ions i.e. Na+/Cl- to neutralize the system. The physiological salt concentration was applied by addition of 150 millimolar NaCl. The system was energy minimized again after solvation. OPLS 2005 Force Field [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] was used for simulation. Constant pressure of 1 atm bar and temperature of 310 Kelvin were used using Berendsen Barostat to run the simulation. The time step of MD iteration was set to 2 femto-second using the SHAKE algorithm. To calculate the long range iteration Particle Mesh Ewald method was used. MD simulation of 1 microsecond time was calculated and recorded in the trajectory i.e., frame after every 1 nanosecond. The trajectory output was saved in .DTR and .CMS file format to store the trajectory and structure data respectively.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Dimension Reduction Techniques\u003c/h2\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003e2.2.1 T-sne (t-Distributed Stochastic Neighbor Embedding)\u003c/h2\u003e \u003cp\u003eT-sne (t-Distributed Stochastic Neighbor Embedding) [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] is the nonlinear dimensionality reduction technique that effectively projects the high dimensional data into low dimension and preserve the original data as much as possible. It used to embed high dimensional data into low dimensional data. T-sne is the variation of Stochastic Neighbor Embedding (sne) in which the similarity between the sample \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e for P-dimensional samples \u003cem\u003ex\u003c/em\u003e\u003csub\u003e1\u003c/sub\u003e, \u003cem\u003ex\u003c/em\u003e\u003csub\u003e2\u003c/sub\u003e, \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003e3\u003c/em\u003e\u003c/sub\u003e, .. \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003en\u003c/em\u003e\u003c/sub\u003e is represented by finding out the conditional probability \u003cem\u003eP\u003c/em\u003e\u003csub\u003e\u003cem\u003ei|j\u003c/em\u003e,\u003c/sub\u003e where, \u003cem\u003ei,j\u0026thinsp;=\u0026thinsp;1, 2, 3, ...n. P\u003c/em\u003e\u003csub\u003e\u003cem\u003ei|j\u003c/em\u003e\u003c/sub\u003e is high for the samples having high similarity but \u003cem\u003eP\u003c/em\u003e\u003csub\u003e\u003cem\u003ei|j\u003c/em\u003e\u003c/sub\u003e is almost zero for widely separated samples. The \u003cem\u003ep\u003c/em\u003e\u003csub\u003e\u003cem\u003ej|i\u003c/em\u003e\u003c/sub\u003e is given as,\u003c/p\u003e \u003cp\u003e \u003cem\u003ep\u003c/em\u003e \u003csub\u003e \u003cem\u003ej|i\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e=exp(\u0026minus;\u0026thinsp;d\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e(x\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e)/2\u003c/em\u003e\u003csub\u003e\u003cem\u003eσ\u003c/em\u003e\u003c/sub\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e)/\u0026sum;\u003c/em\u003e\u003csub\u003e\u003cem\u003ek 6\u0026ne;i\u003c/em\u003e\u003c/sub\u003e \u003cem\u003eexp(\u0026minus;\u0026thinsp;d\u003c/em\u003e\u003csub\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e(x\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, \u003cem\u003ex\u003c/em\u003e\u003csub\u003e\u003cem\u003ek\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e)/2\u003c/em\u003e\u003csub\u003e\u003cem\u003eσ\u003c/em\u003e\u003c/sub\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e)\u003c/em\u003e\u003c/p\u003e \u003cp\u003e \u003cem\u003efor i\u0026thinsp;\u0026ne;\u0026thinsp;j, and p\u003c/em\u003e \u003csub\u003e \u003cem\u003ei|i\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e= 0\u003c/em\u003e,\u003c/p\u003e \u003cp\u003eWhere \u003cem\u003ed\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e (\u003cb\u003ex\u003c/b\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, \u003cb\u003ex\u003c/b\u003e \u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e) represents the square of the Euclidean distance between the sample \u003cb\u003ex\u003c/b\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and sample \u003cb\u003ex\u003c/b\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e and \u003csub\u003eσ\u003c/sub\u003e\u003csup\u003e2\u003c/sup\u003e\u003csub\u003ei\u003c/sub\u003e represents the variance of the Gaussian distribution, centred on sample \u003cb\u003ex\u003c/b\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003eProbability in the original space is defined mathematically as;\u003c/p\u003e \u003cp\u003e \u003cem\u003ep\u003c/em\u003e \u003csub\u003e \u003cem\u003ej|i =(\u003c/em\u003e \u003c/sub\u003e \u003cem\u003ep\u003c/em\u003e \u003csub\u003e \u003cem\u003ej|i +\u003c/em\u003e \u003c/sub\u003e \u003cem\u003ep\u003c/em\u003e \u003csub\u003e \u003cem\u003ei|j )/ 2n\u003c/em\u003e \u003c/sub\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003efor i\u0026thinsp;\u0026ne;\u0026thinsp;j, and p\u003c/em\u003e \u003csub\u003e \u003cem\u003eii\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e= 0\u003c/em\u003e,\u003c/p\u003e \u003cp\u003eWhere n is the size of the sample. In the t-SNE algorithm \u0026ldquo;perplexity\u0026rdquo; is an input parameter that specifies the number of neighbours. It can be mathematically defined as\u003c/p\u003e \u003cp\u003e \u003cem\u003eperp(P\u003c/em\u003e \u003csub\u003e \u003cem\u003ei\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e)\u0026thinsp;=\u0026thinsp;2\u003c/em\u003e \u003csup\u003e \u003cem\u003eH(Pi)\u003c/em\u003e \u003c/sup\u003e \u003c/p\u003e \u003cp\u003eWhere H(Pi) is the Shannon entropy.\u003c/p\u003e \u003cp\u003eT-sne employs student t- distribution (with single degree of freedom) to avoid overcrowding.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.2.2 Isomap\u003c/h2\u003e \u003cp\u003eIsomap is the nonlinear dimensionality reduction technique that preserves the geodesic proximities using non-euclidean distances [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The geodesic metric precisely preserves the inter-point distances if the data lies in the nonlinear manifold. The isomap algorithm works in three steps. In the first step inter-point distance x\u003csub\u003eij\u003c/sub\u003e is constructed for neighbouring points. This is done by considering K-nearest points for the target point. Then the weighted graph is used to represent the distance x\u003csub\u003eij\u003c/sub\u003e, where weight of the edge defines the distance for all i and j. The second step is to define the geodesic distance g\u003csub\u003eij\u003c/sub\u003e for two points by using shortest path in the graph. The dissimilarity matrix \u003cem\u003eΔ\u003c/em\u003e =(g\u003csub\u003eij\u003c/sub\u003e ) is constructed then the last step of isomap is done by applying multidimensional scaling(MDS) to find out the embedding space and associated principal components.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Clustering\u003c/h2\u003e \u003cp\u003eClustering is the unsupervised learning technique to partition the data without any prior knowledge of the data and it provides more insight about the complex dataset. In clustering the identical groups or partitions are identified and grouped into the same group [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Different types of clustering algorithms are; Hierarchical clustering, Partition clustering and density based clustering. In hierarchical clustering the hierarchy of clusters is formed based on the similarity in the data. The hierarchy of data is presented in the form of a tree called dendrogram. Two approaches of hierarchical clustering are; agglomerative and divisive. In agglomerative every data point is considered as a single cluster and then the clusters are merged based on some similarity index. This process continues until one cluster is formed. The divisive algorithm is the reverse of the agglomerative approach. In partition clustering the data points are partitioned into a user-specified number of clusters (K) based on some similarities among data. Some of the popular partition clustering algorithms are; K-Means [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], PAM (K-Medoids)[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], CLARA (Clustering Large Applications)[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] etc.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Validation Index\u003c/h2\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.4.1 Average Silhouette Width\u003c/h2\u003e \u003cp\u003eAverage Silhouette Width (ASW) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] is the internal validation index used to find out the optimal number of clusters. It is obtained by the difference between average distance between within the cluster and minimum distance between the clusters. The number of clusters for which ASW is maximum, is the optimal number of clusters. Mathematically calculated as:\u003c/p\u003e \u003cp\u003e \u003cem\u003e\u0026#119878; = 1/ (( \u0026sum;\u003c/em\u003e \u003csup\u003e \u003cem\u003en\u003c/em\u003e \u003c/sup\u003e \u003csub\u003e \u003cem\u003ei=1\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e( \u0026#119887;(\u0026#119894;) \u0026minus; \u0026#119886;(\u0026#119894;))/ (\u0026#119898;\u0026#119886;\u0026#119909;{\u0026#119886;(\u0026#119894;), \u0026#119887;(\u0026#119894;)}))\u003c/em\u003e \u003c/p\u003e \u003cp\u003eWhere, a(i) represents the average distance within the cluster, b(i) represents the minimum distance between the clusters. The number of clusters for which ASW is maximum is the optimal number of clusters\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.4.2 Davies Bouldin Index\u003c/h2\u003e \u003cp\u003eThe Davies\u0026ndash;Bouldin(DB) index[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] is calculated as the ratio between \u0026ldquo;within-cluster\u0026rdquo; and \u0026ldquo;between-cluster\u0026rdquo; distances. The number of clusters for which DB index is minimum, is the optimal number of clusters. Mathematically calculated as:\u003c/p\u003e \u003cp\u003e \u003cem\u003eDB(C)\u0026thinsp;=\u0026thinsp;1/k(\u0026sum;\u003c/em\u003e \u003csup\u003e \u003cem\u003ek\u003c/em\u003e \u003c/sup\u003e \u003csub\u003e \u003cem\u003ei=1\u003c/em\u003e \u003c/sub\u003e \u003cem\u003emax\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u0026le;k,j\u0026ne;i\u003c/em\u003e\u003c/sub\u003e\u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003eij\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e), k=|c|\u003c/em\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003eij\u003c/em\u003e\u003c/sub\u003e is the \"ratio of within-to-between cluster distance\" for the i\u003csup\u003eth\u003c/sup\u003e and j\u003csup\u003eth\u003c/sup\u003e clusters. \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003eij\u003c/em\u003e\u003c/sub\u003e \u003cem\u003eis\u003c/em\u003e mathematically calculated as:\u003c/p\u003e \u003cp\u003e \u003cem\u003eC\u003c/em\u003e \u003csub\u003e \u003cem\u003eij\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e=(C\u0026rsquo;\u003c/em\u003e \u003csub\u003e \u003cem\u003ei\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e+C\u0026rsquo;\u003c/em\u003e \u003csub\u003e \u003cem\u003ej\u003c/em\u003e \u003c/sub\u003e \u003cem\u003e)/C\u003c/em\u003e \u003csub\u003e \u003cem\u003eij\u003c/em\u003e \u003c/sub\u003e \u003c/p\u003e \u003cp\u003eWhere, \u003cem\u003eC\u0026rsquo;\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eC\u0026rsquo;\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e is the average distance between every point in clusters and their centroids. \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003eij\u003c/em\u003e\u003c/sub\u003e is the distance (Euclidean distance) between the two cluster centroids.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.4.3 Dunn Index\u003c/h2\u003e \u003cp\u003eDunn Index[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] is used to identify the well separated, compact clusters and small variance clusters. Higher the value of the Dunn index the better is the clustering technique. It can be calculated as:\u003c/p\u003e \u003cp\u003e \u003cem\u003eDunn index (U)\u0026thinsp;=\u0026thinsp;min\u003c/em\u003e \u003csub\u003e\u003cem\u003e1\u0026le;i\u0026le;c\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e{min\u003c/em\u003e \u003csub\u003e\u003cem\u003e1\u0026le;j\u0026le;c, j\u0026ne;i\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e{(\u0026#120575;( C\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e))/ (max\u003c/em\u003e \u003csub\u003e\u003cem\u003e1\u0026le;k\u0026le;c\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e(\u0026#120575;(C\u003c/em\u003e\u003csub\u003e\u003cem\u003ek\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e))}}\u003c/em\u003e\u003c/p\u003e \u003cp\u003eWhere, (\u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e, \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e ) is the inter-cluster distance between \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ei\u003c/em\u003e\u003c/sub\u003e and \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e. \u0026#120575;(\u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ek\u003c/em\u003e\u003c/sub\u003e) is the intra-cluster distance within cluster \u003cem\u003eC\u003c/em\u003e\u003csub\u003e\u003cem\u003ek\u003c/em\u003e\u003c/sub\u003e.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. Result and Discussion","content":"\u003cp\u003eX-Ray Crystallographic structure of the protein provides details of its 3D architecture and information about its binding pocket, as such insights are static in nature, and the dynamic nature of the protein cannot be understood well through these methods alone. Besides, dynamic behaviour of the protein plays a vital role in the drug discovery process. It can be well understood by MD simulation. In this study, 1 microsecond long MD simulation of CDK2 Kinase done and 1000 conformations were generated using the Desmond MD tool.\u003c/p\u003e\n\u003cp\u003eAll the frames of the MD trajectory were aligned to null down the translations. The alpha carbon atoms of each of the amino acids were selected as the representative atom for the root mean square deviation (RMSD) analysis as shown in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. Through the RMSD plots it was observed that the mean deviation was ~\u0026thinsp;3.5 angstroms, N lobe amino acids deviated more than the C lobe amino acids. This suggests that there was considerable deviation in the kinase structure during the course of the MD. This prompted us to view the contribution of each amino acid in the fluctuation during the MD using the root mean square fluctuation (RMSF) analysis. The RMSF of the Cdk2 is shown in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. The pattern of the local fluctuation profile of the amino acids was in sync with the already published literature [\u003cspan class=\"CitationRef\"\u003e19\u003c/span\u003e]. The loops recorded more deviations; the N terminal residues of the C helix also had relatively higher fluctuation [\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e]. The most fluctuations were observed in amino acids mapped on activation loop and the GHI sub helical domain. The E and F helix had least fluctuations.\u003c/p\u003e\n\u003cp\u003eFollowing the deviation and fluctuational analysis, T-sne and Isomap were applied as dimensional reduction techniques to reduce the dimensionality of data. Then different clustering techniques; hierarchical, K-Means, HK-Means, PAM and CLARA were applied to cluster the data. Different indexes (Average Silhouette Width, Davies Bouldin Index and Dunn Index) were used to find out the best clustering algorithm and optimal number of clusters. Values of different indexes were recorded for five clustering techniques by varying the number of clusters from 2 to 10.\u003c/p\u003e\n\u003cp\u003eT-sne was used for dimensionality reduction and different clustering techniques are applied to cluster the data. The value of Average Silhouette Width (ASW), Davies Bouldin(DB) Index and Dunn Index for different clustering techniques were recorded as shown in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e and Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e respectively. It was observed from Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, that all the clustering techniques have high ASW value when 10 clusters were formed. Out of 5 clustering techniques Hierarchical K-Means clustering recorded the highest value of 0.512444 (for 10 clusters). Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e shows the value DB index for different clustering techniques; hierarchical, K-Means, PAM and ClARA clustering with DB value of 0.690332, 0.6987556, 0.7072587 and 0.700863 respectively, showed 9 as the optimal number of clusters except HK-Means with DB value of 0,696726 showed 10 as optimal number of clusters.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShow the values of Average Silhouette Width (ASW) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.3885\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.3818612\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4014626\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.388876\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.389238\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.3962\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4463395\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4149279\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.465571\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.435407\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4297\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4595713\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4469391\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.459571\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.447717\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4654\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4878133\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4785397\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.482815\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.434354\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4831\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4781707\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4763377\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.491696\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.465199\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4861\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4629503\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4937016\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.495461\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.478229\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4848\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.476702\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4758134\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.487784\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.465917\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4958\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4958097\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.490095\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.49581\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.481096\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5005\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.4997089\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5051389\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.512444\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.50783\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShows the values of Davies Bouldin index (DB) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM\u003c/p\u003e\n \u003cp\u003eClustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHK-Means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara\u003c/p\u003e\n \u003cp\u003eClustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.1258234\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.2367192\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.2049045\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.19084\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e1.222572\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.8347396\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.8284007\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.9258379\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.806557\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.848357\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.8051623\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.725148\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7660555\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.725148\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.721648\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7605432\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7083599\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7115981\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.702764\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.743582\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7517587\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7154494\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7234466\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.708506\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.741091\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7643028\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7899481\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7079546\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.726907\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.74704\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7707523\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.722547\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7272507\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.725646\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.732795\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.690332\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6987556\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7072587\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.698756\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.700863\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7437819\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7080912\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.7155078\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.696726\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.700888\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShow the values of Dunn Index (DI) for different clustering techniques by varying the number of clusters from 2 to 10 when T-sne is used for dimensionality reduction\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0243\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0096\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0101\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.00935\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0036\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0311\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0087\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0112\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.008696\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0072\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.032\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0057\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0066\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.005683\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0046\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.011\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0216\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0092\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.006271\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0026\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0136\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0153\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0061\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.015328\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0099\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0136\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0113\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0077\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.01126\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0112\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0136\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.016\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0113\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.01126\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0083\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0136\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0184\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0113\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.008241\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0034\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0188\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0184\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0089\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.010273\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0091\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eValue of the Dunn index is shown in Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. Hierarchical, K-Means, PAM and ClARA clustering reported 0.032, 0.0216, 0.0113, 0.015328 and 0.0112 respectively as optimal value of Dunn index for 4, 5, 9, 6 and 7 respectively as the optimal number of clusters.\u003c/p\u003e\n\u003cp\u003eWhen Isomap was used for dimensionality reduction and different clustering techniques are applied to cluster the data. The value of ASW for different clustering techniques was shown in Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e. It was observed that all the clustering techniques have high ASW value when 5 clusters are formed. Out of five clustering techniques; K-Means and Hierarchical K-Means clustering have recorded the highest value of 0.604545 (for 5 clusters). Table \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e shows that all the clustering techniques; hierarchical, K-Means, PAM, Hierarchical K-Means and CLARA clustering with DB value of 0.5996, 0.6045447, 0.6040297, 0.604545 and 0.599434 resp. show 5 as the optimal number of clusters. Table \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e shows that the value of DI for hierarchical clustering, K-Means, PAM Hierarchical K-Means and CLARA clustering is 0.047, 0.0571, 0.0236, 0.046064 and 0.0089 resp. for 10,7,8,10 and 3 resp. as the optimal number of clusters. It was observed that ASW, DB and DI had shown better values for different clustering techniques when Isomap was used as dimensional reduction techniques so for finding the optimal numbers of clusters Isomap was considered for dimensional reductionality. It was observed that K-Means clustering shows better index values (ASW and DB) for five as an optimal number of clusters. So we have considered five as the optimal number of clusters.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShow the values of Average Silhouette Width (ASW) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5996\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5997813\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5962576\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.599781\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.594198\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5286\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5777035\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5802975\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5783\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.580945\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5325\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5864056\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5899616\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.586406\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.589118\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5919\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6045447\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6040297\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.604545\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.599434\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5521\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5648679\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5692951\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.564868\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.555593\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5324\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5438552\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5313613\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.554465\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.528342\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5329\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5234979\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5314133\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.550032\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.512257\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5184\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5223692\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5339954\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.528485\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.508167\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5049\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5085656\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5172938\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.495137\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.494295\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable id=\"Tab5\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShows the values of Davies Bouldin index (DB) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHk-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6324226\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6306561\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6293577\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.630656\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.627538\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6024295\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5863843\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5747294\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.58534\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.571271\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5526272\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5764032\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.55947\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.576403\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.56181\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.546069\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5450825\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5462048\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.545083\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.545521\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6005217\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5959664\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5877089\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.595966\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.605961\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5929847\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6129836\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6154011\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.603058\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.624783\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.5956353\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6551261\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6121824\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.615468\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.645534\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6129884\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.639297\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6228808\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6345\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.642229\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.624127\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6810369\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.6215201\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.673057\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.637879\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable id=\"Tab6\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eShow the values of Dunn Index (DI) for different clustering techniques by varying the number of clusters from 2 to 10 when Isomap is used for dimensionality reduction.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHierarchical Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePAM Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHK-means Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eClara Clustering\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0177\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0034\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0183\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.003439\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0006\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0168\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0047\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0174\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.016919\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0089\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.028\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0077\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.025194\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0034\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0291\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0397\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0079\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.039718\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0029\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0375\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0438\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0071\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.040631\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0058\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0376\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0571\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0071\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.008078\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0068\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0422\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0446\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0236\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.044116\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0058\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0447\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.045\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.007\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.040499\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0038\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 Clusters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.047\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0567\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0098\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.046064\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0036\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eFor docking five frame representatives were created by finding out the average structure generated by computing the average mean of the Cartesian coordinates of each amino acid in all the structures present in a cluster. Then this average structure of each cluster was further used to identify the representative for each cluster. And the Structure having minimum deviation with average structure was found and saved as frame representative for each cluster. The representations thus generated were used in ensemble docking. The frame representatives thus generated were docked with drug-like molecules taken from 52,000 compounds of Maybridge Library ((\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.maybridge\u003c/span\u003e\u003c/span\u003e. com)) using molegro virtual docker software [\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e]. The ranking of the drug molecule was done on the basis of docking energies derived by the physico-chemical interactions, shape complementarity and entropy. The top 6 compounds with high docking scores were listed in Table\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e. All these compounds cleared the Lipinski\u0026rsquo;s filter and had molecular weight between 334.4 and 463. Their rotatable bonds were in the range of 3 to 5 and that were checked using the Swiss ADME tool [\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e] for the ADME properties that is shown in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab7\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eThe top 6 compounds with strong binding energies in terms of Molecular Docking Score was shown\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eS. No\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eZINC compound ID rotatable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eNumber of rotatable bonds\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMolDock Score\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMolecular weight (g/ mol)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIUPAC name\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC12369719\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-136.131\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e376.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC04394657\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-135.565\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e394.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eN-[(2-chlorophenyl)methylideneamino]-4-(5-methyl-3-phenyl-1,2-oxazol-4-yl)-1,3-thiazol-2-amine\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC01047457\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-135.456\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e398.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2-(3,4-dichlorophenyl)-\u003cem\u003eN\u003c/em\u003e\u0026apos;-(thiophene-2-carbonyl)-1,3-thiazole-4-carbohydrazide\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC00169253\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-132.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e334.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4-[2-(2,1,3-benzoxadiazol-5-yloxymethyl)-1,3-thiazol-4-yl]benzonitrile\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC13658326\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-130.353\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e463\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4-[(Z)-[3-(4-chlorophenyl)-1-phenylpyrazol-4-yl]methylideneamino]-3-thiophen-2-yl-1H-1,2,4-triazole-5-thione\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eZINC12369719\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-127.131\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e376.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"},{"header":"4. Conclusion","content":"\u003cp\u003eIn the present study, molecular dynamic simulation of CDK2 was done and its diverse 3D conformations were generated. T-sne and Isomap were used as dimensional reduction techniques and then different clustering algorithms including Hierarchical, K-means, PAM, Hierarchical K-means and Clara were applied. Then validation indexes including Average Silhouette Width, Dunn Index and Davies Bouldin Index were used to evaluate dimensional reduction and different clustering techniques thus applied. Then the optimal number of clusters is found out, to find out the representative structure required for docking. Then the representative structures are virtually screened with the Maybridge database. Six novel target molecules were identified with promising binding energies and their ADMET properties were evaluated. Top 6 selected compounds: ZINC12369719, ZINC4394657, ZINC1047457, ZINC0016925, ZINC13658326 and ZINC12369719. The compounds thus found have promising scope for wet lab screening and design of inhibitors against CDK2.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e R.K, I.S, J.M, Va.S and Vi.S conceived the idea. R.K, I.S and Va. S executed the experiment and wrote the article. J.M and Vi.S did edition and corrections.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eNo funding\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e The authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor’s Biography\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRajneet Kaur Bijral\u003c/strong\u003e completed her Ph.D in Computer Science from Department of Computer Science and IT, University of Jammu. She has an experience of over 5 years of teaching experience. \u0026nbsp;Her area of interest includes Artificial intelligence, machine learning, deep learning, Bioinfomatic etc.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDr. Inderpal Singh\u003c/strong\u003e is a reputed researcher in the field of Bioinfomatics. His area of research includes cancer drug discovery, kinase structure, genetics of rare disorder, metagenomics and immuoinformatics. He completed his Ph.D in Biotechnology from SMVDU University and served as faculty in SKUAST before starting his research company Bioinfores where he teaches biological data science analysis to researcher and professionals across globe. His publications are published across most reputed international publication house.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDr. Varun Sharma\u003c/strong\u003e is Head of Bioinformatics at NMC Genetics India Pvt. Ltd.(NMC Healthcare). He has done his Ph.D. from Shri Mata Vaishno Devi University and Postdoc from Birbal Sahni Institute of Paleosciences, Lucknow, GoI. His research work includes pharmacogenomics evaluation of different complex disorders. He has been awarded with young scientist fellowship for his work on Pharmacogenomics by Council of Lindau Nobel Laureates Germany. He is having more than 30 research articles in the peer reviewed journals and book chapters. With an experience of more than 5 years he has great hold on developing statistical models for different diseases in correlation with their phenotypes, Genome wide association analyses and functional mapping and annotation of genetic association.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDr. Jatinder Manhas\u003c/strong\u003e completed his MCA and Ph.D. in Computer Science from Department of Computer Science and IT, University of Jammu. He is an eminent Researcher and his\u0026nbsp;publications are published across most reputed international publication house. \u0026nbsp;He has a vast experience of over 13 years in the field of networking, databases, and website design issues. Currently he is working as a Senior Assistant Professor in the Department of Computer Science and IT, Bhaderwah Campus of University of Jammu. His area of interest includes Artificial intelligence, machine learning, deep learning, medical image analysis, website design issues, etc.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDr. Vinod Sharma\u0026nbsp;\u003c/strong\u003ereceived his Ph.D. in Computer Science from Department of Computer Science and IT, University of Jammu. He is an eminent Researcher and his\u0026nbsp;publications are published across most reputed international publication house. \u0026nbsp;He has over 26 years of experience in teaching and research. Currently he is working as a Professor in the Department of Computer Science and IT, University of Jammu. Besides this, he is also appointed as Director, Ramnagar Campus, and University of Jammu. His area of interest includes Artificial intelligence, machine learning, deep learning, medical image analysis, medical diagnosis, etc.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eS. Hu, \u0026nbsp; A. V. Danilov, K. Godek, B. Orr, L. J. Tafe., J. Rodriguez-Canales, \u0026nbsp;C. Behrens, B. Mino, C. A. Moran, V. A. Memoli, L. M. Mustachio, F. Galimberti, S. Ravi, A. DeCastro, Y. Lu, D. Sekula, A. S. Andrew, \u0026nbsp;I. I. Wistuba, S. Freemantle, D. A. Compton, \u0026nbsp;\u0026hellip; Dmitrovsky, E. (2015). \u003cem\u003e\u0026ldquo;CDK2 Inhibition Causes Anaphase Catastrophe in Lung Cancer through the Centrosomal Protein CP110,\u0026rdquo;\u003c/em\u003e\u003cem\u003eCancer research\u003c/em\u003e, vol \u003cem\u003e75\u003c/em\u003e, no 10, pp. 2029\u0026ndash;38. doi.org/10.1158/0008-5472.Can-14-1494\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eS. Akli, C. S. Van Pelt, T. Bui, L. Meijer, and K. Keyomarsi, \u0026quot;Cdk2 is required for breast cancer mediated by the low-molecular-weight isoform of cyclin E,\u0026quot; (in eng), \u003cem\u003eCancer Res,\u0026nbsp;\u003c/em\u003evol. 71, no. 9, pp. 3377-86, May 1 2011, doi: 10.1158/0008-5472.Can-10-4086.\u003c/li\u003e\n \u003cli\u003eJ. D. Durrant and J. A. McCammon, \u0026quot;Molecular dynamics simulations and drug discovery,\u0026quot; \u003cem\u003eBMC Biology,\u0026nbsp;\u003c/em\u003evol. 9, no. 1, p. 71, 2011/10/28 2011, doi: 10.1186/1741-7007-9-71.\u003c/li\u003e\n \u003cli\u003eA. Hospital, J. R. Go\u0026ntilde;i, M. Orozco, and J. L. Gelp\u0026iacute;, \u0026quot;Molecular dynamics simulations: advances and applications,\u0026quot; (in eng), \u003cem\u003eAdv Appl Bioinform Chem,\u0026nbsp;\u003c/em\u003evol. 8, pp. 37-47, 2015, doi: 10.2147/aabc.S70333.\u003c/li\u003e\n \u003cli\u003eG. Sliwoski, S. Kothiwale, J. Meiler, and E. W. Lowe, Jr., \u0026quot;Computational methods in drug discovery,\u0026quot; (in eng), \u003cem\u003ePharmacol Rev,\u0026nbsp;\u003c/em\u003evol. 66, no. 1, pp. 334-95, 2014, doi: 10.1124/pr.112.007336.\u003c/li\u003e\n \u003cli\u003eH. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig, I. N. Shindyalov, P. E. Bourne,\u0026nbsp;\u0026quot;The Protein Data Bank,\u0026quot; (in eng), \u003cem\u003eNucleic Acids Res,\u0026nbsp;\u003c/em\u003evol. 28, no. 1, pp. 235-42, Jan 1 2000, doi: 10.1093/nar/28.1.235.\u003c/li\u003e\n \u003cli\u003eM. Rostkowski, M. H. Olsson, C. R. S\u0026oslash;ndergaard, and J. H. Jensen, \u0026quot;Graphical analysis of pH-dependent properties of proteins predicted using PROPKA,\u0026quot; (in eng), \u003cem\u003eBMC Struct Biol,\u0026nbsp;\u003c/em\u003evol. 11, p. 6, Jan 26 2011, doi: 10.1186/1472-6807-11-6.\u003c/li\u003e\n \u003cli\u003eM. Harrach and B. Drossel, \u0026quot;Structure and dynamics of TIP3P, TIP4P, and TIP5P water near smooth and atomistic walls of different hydroaffinity,\u0026quot; \u003cem\u003eThe Journal of chemical physics,\u0026nbsp;\u003c/em\u003evol. 140, p. 174501, 05/07 2014, doi: 10.1063/1.4872239.\u003c/li\u003e\n \u003cli\u003eD. Shivakumar, E. Harder, W. Damm, R. Friesner, and W. Sherman, \u0026quot;Improving the Prediction of Absolute Solvation Free Energies Using the Next Generation OPLS Force Field,\u0026quot; \u003cem\u003eJournal of Chemical Theory and Computation,\u0026nbsp;\u003c/em\u003evol. 8, pp. 2553\u0026ndash;2558, 07/09 2012, doi: 10.1021/ct300203w.\u003c/li\u003e\n \u003cli\u003eD. Kobak and P. Berens, \u0026quot;The art of using t-SNE for single-cell transcriptomics,\u0026quot; \u003cem\u003eNature Communications,\u0026nbsp;\u003c/em\u003evol. 10, no. 1, p. 5416, 2019/11/28 2019, doi: 10.1038/s41467-019-13056-x.\u003c/li\u003e\n \u003cli\u003eC. Wang, W. Fu, H. Huang, and J. Chen, \u0026quot;Isomap-Based Three-Dimensional Operational Modal Analysis,\u0026quot; \u003cem\u003eScientific Programming,\u0026nbsp;\u003c/em\u003evol. 2020, p. 6348372, 2020/07/14 2020, doi: 10.1155/2020/6348372.\u003c/li\u003e\n \u003cli\u003eC. Maione, D. R. Nelson, and R. M. Barbosa, \u0026quot;Research on social data by means of cluster analysis,\u0026quot; \u003cem\u003eApplied Computing and Informatics,\u0026nbsp;\u003c/em\u003evol. 15, no. 2, pp. 153-162, 2019/07/01/ 2019, doi:\u0026nbsp;\u003ca href=\"https://doi.org/10.1016/j.aci.2018.02.003\"\u003ehttps://doi.org/10.1016/j.aci.2018.02.003\u003c/a\u003e.\u003c/li\u003e\n \u003cli\u003eS. Na, L. Xumin, and G. Yong, \u0026quot;Research on k-means Clustering Algorithm: An Improved k-means Clustering Algorithm,\u0026quot; in \u003cem\u003e2010 Third International Symposium on Intelligent Information Technology and Security Informatics\u003c/em\u003e, 2-4 April 2010 2010, pp. 63-67, doi: 10.1109/IITSI.2010.74.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eD. Lei, Q.-s. Zhu, J. Chen, H. Lin, and P. Yang, \u0026quot;Automatic PAM Clustering Algorithm for Outlier Detection,\u0026quot; \u003cem\u003eJournal of Software,\u0026nbsp;\u003c/em\u003evol. 7, 04/27 2012, doi: 10.4304/jsw.7.5.1045-1051.\u003c/li\u003e\n \u003cli\u003eT. Gupta and S. P. Panda, \u0026quot;Clustering Validation of CLARA and K-Means Using Silhouette \u0026amp; DUNN Measures on Iris Dataset,\u0026quot; in \u003cem\u003e2019 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (COMITCon)\u003c/em\u003e, 14-16 Feb. 2019 2019, pp. 10-13, doi: 10.1109/COMITCon.2019.8862199.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eF. Batool and C. Hennig, \u0026quot;Clustering with the Average Silhouette Width,\u0026quot; \u003cem\u003eComputational Statistics \u0026amp; Data Analysis,\u0026nbsp;\u003c/em\u003evol. 158, p. 107190, 02/01 2021, doi: 10.1016/j.csda.2021.107190.\u003c/li\u003e\n \u003cli\u003eJ. Xiao, J. Lu, and X. Li, \u0026quot;Davies Bouldin Index based hierarchical initialization K-means,\u0026quot; \u003cem\u003eIntelligent Data Analysis,\u0026nbsp;\u003c/em\u003evol. 21, pp. 1327-1338, 11/15 2017, doi: 10.3233/IDA-163129.\u003c/li\u003e\n \u003cli\u003eJ. C. Bezdek and N. R. Pal, \u0026quot;Some new indexes of cluster validity,\u0026quot; \u003cem\u003eIEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics),\u0026nbsp;\u003c/em\u003evol. 28, no. 3, pp. 301-315, 1998, doi: 10.1109/3477.678624.\u003c/li\u003e\n \u003cli\u003eT. Mohammad, S. Batra, R. Dahiya, MH. Baig, IA. Rather, JJ. Dong, I. Hassan, \u0026quot;Identification of High-Affinity Inhibitors of Cyclin-Dependent Kinase 2 Towards Anticancer Therapy,\u0026quot; (in eng), \u003cem\u003eMolecules,\u0026nbsp;\u003c/em\u003evol. 24, no. 24, Dec 15 2019, doi: 10.3390/molecules24244589.\u003c/li\u003e\n \u003cli\u003eK. Huang, Y. H. Wang, A. Brown, and G. Sun, \u0026quot;Identification of N-terminal lobe motifs that determine the kinase activity of the catalytic domains and regulatory strategies of Src and Csk protein tyrosine kinases,\u0026quot; (in eng), \u003cem\u003eJ Mol Biol,\u0026nbsp;\u003c/em\u003evol. 386, no. 4, pp. 1066-77, Mar 6 2009, doi: 10.1016/j.jmb.2009.01.012.\u003c/li\u003e\n \u003cli\u003eG. Bitencourt-Ferreira and W. F. de Azevedo, Jr., \u0026quot;Molegro Virtual Docker for Docking,\u0026quot; (in eng), \u003cem\u003eMethods Mol Biol,\u0026nbsp;\u003c/em\u003evol. 2053, pp. 149-167, 2019, doi: 10.1007/978-1-4939-9752-7_10.\u003c/li\u003e\n \u003cli\u003eA. Daina, O. Michielin, and V. Zoete, \u0026quot;SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules,\u0026quot; \u003cem\u003eScientific Reports,\u0026nbsp;\u003c/em\u003evol. 7, no. 1, p. 42717, 2017/03/03 2017, doi: 10.1038/srep42717.\u003c/li\u003e\n \u003cli\u003eD. P. Singh, G. Abhishek, and K. Baijnath, \u0026quot;DWUT-MLP: Classification of anticancer drug response using various feature selection and classification techniques,\u0026quot; \u003cem\u003eChemometrics and Intelligent Laboratory Systems,\u0026nbsp;\u003c/em\u003evol. 225, p. 104562, 2022/06/15/ 2022, doi: https://doi.org/10.1016/j.chemolab.2022.104562.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"structural-chemistry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"stuc","sideBox":"Learn more about [Structural Chemistry](https://www.springer.com/journal/11224)","snPcode":"11224","submissionUrl":"https://submission.nature.com/new-submission/11224/3","title":"Structural Chemistry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"CDK2 kinase, Isomap, T-sne, Clustering, Simulation","lastPublishedDoi":"10.21203/rs.3.rs-4315752/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4315752/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDeregulation of Cyclin-Dependent kinase 2 is one of the main causes of lung cancer and is considered as a potential target for cancer therapy. Discovery of the inhibitors against the kinase using traditional methods is costly and time consuming. This issue can be easily dealt with by the usage of potential computational methods. Molecular Dynamic simulation has the ability to offer scientists an in-depth understanding of the dynamics of biomolecules at an atomic level. In the present study the molecular dynamic simulation of Cyclin Dependent kinase is carried out for one microsecond. T-sne and Isomap dimensionality reduction techniques are used followed by different clustering algorithms; Hierarchical, K-means, PAM, Hierarchical K-means and Clara. Subsequently, an optimal number of clusters was generated by using the validation index; Average Silhouette Width (ASW), Dunn Index and Davies Bouldin Index. Further, the cluster representative for each cluster is virtually screened with 52,000 drug-like molecules taken from May bridge library. Based on docking score the top 6 drugs like molecules are selected. Finally, the ADMET properties of the top compounds are evaluated using the Swiss ADMET tool. The strong binding energies and ADMET properties suggest that these molecules are considered to be valuable inhibitors and play a vital role in drug discovery.\u003c/p\u003e","manuscriptTitle":"Exploring Computational Methods to Identify Inhibitors for CDK2 Kinase","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-06 08:12:17","doi":"10.21203/rs.3.rs-4315752/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorAssigned","content":"","date":"2024-04-30T05:09:50+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-04-29T13:26:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"Structural Chemistry","date":"2024-04-24T05:56:18+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"structural-chemistry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"stuc","sideBox":"Learn more about [Structural Chemistry](https://www.springer.com/journal/11224)","snPcode":"11224","submissionUrl":"https://submission.nature.com/new-submission/11224/3","title":"Structural Chemistry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"e93afd5a-137d-4dfc-8dbb-7a1a6fe6c3a8","owner":[],"postedDate":"May 6th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2024-05-06T08:12:17+00:00","versionOfRecord":[],"versionCreatedAt":"2024-05-06 08:12:17","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4315752","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4315752","identity":"rs-4315752","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}