Clustering Approaches for Evaluation and Analysis on Formal Gene Expression Cancer Datasets

International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
228
IJRITCC | July 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
Clustering Approaches for Evaluation and Analysis on Formal Gene Expression
Cancer Datasets.
Ramachandro Majji
Dept of CSE, GMRIT
Rajam, Andhra Pradesh, India.
e-mail: rama00565@gmail.com
Ravi Bramaramba
Dept of IT, GIT
GITAM University,
Visakhpatnam, AP, India.
Abstract:- Enormous generation of biological data and the need of analysis of that data led to the generation of the field Bioinformatics. Data
mining is the stream which is used to derive, analyze the data by exploring the hidden patterns of the biological data. Though, data mining can
be used in analyzing biological data such as genomic data, proteomic data here Gene Expression (GE) Data is considered for evaluation. GE is
generated from Microarrays such as DNA and oligo micro arrays. The generated data is analyzed through the clustering techniques of data
mining. This study deals with an implement the basic clustering approach K-Means and various clustering approaches like Hierarchal, Som,
Click and basic fuzzy based clustering approach. Eventually, the comparative study of those approaches which lead to the effective approach of
cluster analysis of GE.The experimental results shows that proposed algorithm achieve a higher clustering accuracy and takes less clustering
time when compared with existing algorithms.
Keywords: K-Means, Gene expression, Self-organizing map & Hierarchical Clustering
__________________________________________________*****_________________________________________________
I. INTRODUCTION
Microarrays have emerged as a widely used technology for
the monitoring of the expression levels of thousands of
genes during various biological[1]
processes and functions.
Extracting the hidden information in this huge volume of
gene[2]
expression data is quite challenging, and therefore the
need for computationally efficient methods to mine GE data
is a thrust area for the research community.
Moreover, it is a fact that, because of the complexity of the
underlying biological processes, GE data attained from
DNA microarray technologies are mostly noisy and have
very high dimensionality. This scenario makes mining[3]
of
such data very tough and challenging for prediction[4]
.
Several data mining techniques have been used to address
the above mentioned challenge and clustering is one of the
most popular tools found capable in analysing the gene [5]
expression data with better accuracy. Clustering [6]
Techniques (CT) identify the inherent natural structures and
the interesting patterns in the dataset [7].
The purpose of clustering GE data is to reveal the natural
structure inherent in the data. A good CT should depend as
little as possible on prior knowledge. CT for gene
expression [8] data should be capable of extracting useful
information from noisy data. GE data are often highly
connected and may have intersecting and embedded
patterns. Therefore, algorithms for gene-based clustering [9]
should be able to handle this situation effectively. Finally,
biologists are not only interested in the clusters of genes, but
also in the relationships among the clusters and their sub-
clusters, and the relationship among the genes within a
cluster.
CT provides some graphical representation of the cluster
structure, which is much favoured by the biologists to
approach the optimal solution, or a set of approximate
solutions to a range of problems in specific areas.
II. LITERATURE REVIEW
This chapter provides overview of research carried out on
clustering algorithm[10]
and their application to several
microarray data reported in literature.
This chapter is broadly divided into two section. First
section deals with research work carried out on several
microarray data and section two deals with research carried
out on different clustering algorithms.
DNA microarrays are high-throughput methods for
analysing complex nucleic acid samples. It makes possible
to measure rapidly, efficiently and accurately the levels of
expression of all genes present in a biological sample. The
application of such methods in diverse experimental
conditions generates lots of data. However, the main
problem with these data occurs while analysing it.
Derivation of meaningful biological information from raw
microarray data is impeded by the complexity and vastness
of the data. To overcome the problem associated with GE
microarray data many statistical methods has been proposed
in recent past. Some important has been explained below:
Biological data are being produced at a phenomenal rate.
For example as of April 2001, the GenBank repository of
nucleic acid sequences contained 1, 15, 46,000 entries and
the SWISSPROT database of protein sequences contained
95,320 entries. On an average, these databases are doubling
in size every 15 months.
The application of self-organizing maps, a type of
mathematical cluster analysis [11]
that is particularly well
suited for recognizing and classification of the features in
complex, multidimensional data. The approved method is
the nation to exhibit and has been implemented in a

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
229
_______________________________________________________________________________________
publicly stature available computer package, gene cluster,
which performs the analytical calculations and provides easy
data visualization.
The main types of data analysis needed to for biomedical
applications include:
Clustering:: Finding new biological classes or refining
existing ones.
Gene Selection:: In mining terms this is a process of
attribute selection, which finds the genes most strongly
related to a particular class.
Classification:: Classifying diseases or predicting outcomes
based on gene expression patterns and perhaps even
identifying the best treatment for given genetic signature.
Microarrays are a revolutionary new technology with great
potential to provide accurate medical diagnostics help find
the right treatment and cure for many diseases and provide a
detailed genome-wide molecular portrait of cellular states.
DNA microarray makes it possible to quickly, efficiently
and accurately measure the relative representation of each
mRNA species in the total cellular mRNA population. A
DNA experiment consists of measurements of the relative
representation of a large number of mRNA species.
Elucidating the patterns hidden in gene expression data
offers a tremendous opportunity for an enhanced
understanding of functional genomics. Many conventional
clustering algorithms have been adapted or directly applied
to gene expression data, and also new algorithms have
recently been proposed specifically aiming at gene
expression data. These clustering algorithm relevant groups
of genes and samples. In this study the gene expression
clustering is divide into gene based clustering and sample
based clustering. It is explained that K-means a partition
based clustering where no of clusters has to be mentioned
earlier, in hierarchical clustering it produces a genogram, in
SOM it requires the grid structure earlier before clustering
Current approaches to clustering gene expression patterns
utilize hierarchical methods or methods that work for
Euclidean distance metrics. A graph theoretic approach is
considered, and made no assumptions on the similarity
function or the number of clusters sought. The cluster
structure is produced directly, without involving an
intermediate tree stage.
III. PROPOSED STATEMENT
CT have proven to be helpful to understand gene function,
gene regulation, cellular processes and subtypes of cells.
Genes with similar expression patterns can be clustered
together with similar cellular functions. Furthermore, co-
expressed genes in the same cluster are likely to be involved
in the same cellular processes, and a strong correlation of
expression patterns between those genes indicates co-
regulation. Searching for common DNA sequences at the
promoter regions of genes within the same cluster allows
regulatory motifs specific to each gene cluster to be
identified and cis-regulatory elements to be proposed. The
inference of regulation through the clustering of gene
expression data also gives rise to hypotheses regarding the
mechanism of the transcriptional regulatory network.
Finally, clustering different samples on the basis of
corresponding expression profiles may reveal sub-cell types
which are hard to identify by traditional morphology-based
approaches.
IV. INTRODUCTION TO MICRO ARRAY
TECHNOLOGY
4.1 MEASURING MRNA LEVELS
Compared with the traditional approach to genomic
research, which has focused on the local examination and
collection of data on single genes, microarray technologies
have now made itpossible tomonitor the expression levels
for tens of thousands of genes in parallel. The two major
types of microarray experiments are the cDNA microarray
and oligonucleotide arrays. Despite differences in the details
of their experiment protocols, both types’ ofexperiments
involve three common basic procedures:
4.1.1Chip manufacture:
A microarray is a small chip, onto which tens of thousands
of DNA molecules are attached in fixed grids. Each grid cell
relates to a DNA sequence.
The Target Preparation, Labelling & Hybridization [12]
:
Typically, two mRNA samples are reverse-transcribed into
cDNA, labelled using either fluorescent dyes or radioactive
isotopic, and then hybridized with the probes on the surface
of the chip.
The scanning process:
Chips are scanned to read the signal intensity that is emitted
from the labelled and hybridized targets. Generally, both
cDNA microarray and oligo chip experiments measure the
expression level for each DNA sequence by the ratio of
signal intensity between the test sample and the control
sample. Therefore, data sets resulting from both methods
share the same biological semantics. In this study, unless
explicitly stated, both the cDNA microarray data[13]
and the
oligo chip as microarray technology and term the
measurements collected via both methods as gene
expression data.
4.2 PRE-PROCESSING OF GE DATA
A microarray experiment typically assesses a large number
of DNA sequences under multiple conditions. These
conditions may be a time series during a biological process
or a collection of different tissue samples. In this, the focus
on the cluster analysis[14]
of GE data without making a
distinction among DNA sequences, which will uniformly be

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
230
_______________________________________________________________________________________
called “genes”.
Similarly and uniformly referred to all kinds of experimental
conditions as “samples” if no confusion will be caused. A
gene expression data set from a microarray experiment can
be represented by a real-valued
Expression-matrix [15]
where the rows ,
form the expression patterns of genes, the
columns , represent the expression
profiles of samples, and each cell is the measured
expression level of gene as described in the figure.
The original GE matrix obtained from a scanning process
contains noise, missing values and systematic variations
arising from the experimental procedure. Data pre-
processing is indispensable before any cluster analysis can
be performed. Some problems of data pre-processing have
themselves become interesting research topics.
The questionnaire are beyond the scope of this survey; an
examination of the problem of missing value estimation and
the problem of data normalization. Furthermore, many
clustering approaches apply one or more of the following
pre-processing procedures:
a) Filtering out genes with expression levels which do
not change significantly across samples;
b) Performing a mathematical or a logarithmic
transformation function of each expression level;
(or)
Standardizing each row of the gene expression
matrix with a mean of zero and a variance of one.
V. CLUSTERING
In this subsection, the concept of clusters and clustering
preceded with the division of the clustering tasks for gene
expression data into three categories according to different
clustering purposes.
5.1 CLUSTERS AND CLUSTERING
Clustering is the process of grouping data objects into a set
of disjoint classes, called clusters; so that objects within a
class have high similarity to each other, while objects in
separate classes are more dissimilar. Clustering is an
example of unsupervised classification. Classification refers
to a procedure that assigns data objects to a set of classes.
Thus, clustering is distinguished from pattern recognition or
the areas of statistics known as discriminant analysis and
decision analysis, which seek to find rules for classifying
objects from a given set of pre-classified objects.
5.2 CATEGORIES OF GE DATA
Currently, a typical microarray experiment contains to
genes, and this number is expected to reach to the
order of . However, the number of samples involved in a
microarray experiment is generally less than 100.
One of the characteristics of gene expression data is that it is
meaningful to cluster both genes and samples. Besides, the
co-expressed genes can be grouped in clusters based on their
expression patterns. In such gene-based clustering, the
genes are treated as the objects, while the samples are the
features.
The samples can be partitioned into homogeneous groups.
Each group may correspond to some particular macroscopic
phenotype, such as clinical syndromes or cancer types such
sample-based clustering regards the samples as the objects
and the genes as the features. The distinction of gene-based
clustering and sample based clustering is based on different
characteristics of clustering tasks for gene expression data.
Some clustering algorithms, such as K-means and
hierarchical approaches, can be used both to group genes
and to partition samples.
Both the gene-based and sample-based clustering
approaches search exclusive and exhaustive partitions of
objects that share the same feature space.
However, current thinking in molecular biology holds that
only a small subset of genes participates in any cellular
process of interest and that a cellular process takes place
only in a subset of the samples. This belief calls for the
subspaceclustering to capture clusters formed by a subset
of genes across a subset of samples. For subspace clustering
algorithms, genes and samples are treated symmetrically, so
that either genes or samples can be regarded as objects or
features. Furthermore, clusters generated through such
algorithms may have different feature spaces.
While a GE matrix can be analysed from different angles,
the gene-based, sample based clustering and subspace
clustering analysis[16]
face very different challenges. Thus,
there is a scope to adopt very different computational
strategies in the three situations.
VI. METHODOLOGY
The outline of our system or simply architecture of our
system is depicted below.
The major components of our system involves

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
231
_______________________________________________________________________________________
 Gene Expression Data,
 K-Means[17]
 Hierarchal,
 SOM [18]
,
 CLICK & Fuzzy based approach
As shown in the outline, the steps followed to implement
are:
o First move on with the gathering of GE which is
thoroughly pre-processed.
o In the second step, give the data into the system
i.e.., platform
o Now the system will load the data
o After loading, run the K-means clustering
algorithm on it.
o By depicting the results of it the performance of the
algorithm will be noticed
o The above two steps will be performed for
algorithms Hierarchal & SOM.
o Now using the Cluster 3.0 all the above executed
algorithms will be executed in it.
o Later a comparative study [19]
among them will be
performed.
o Now an advanced clustering approach has to kept
for study. For that, Expander tool is used to
implement CLICK clustering algorithm on the
considered data.
o To incorporate a new approach, a collaborative
technique has to be implemented. For that, a
machine learning[20]
concept fuzzy is
collaboratively used with K-means for the scope of
better performance.
o Eventually all the algorithms will be kept under for
a comparative study.
o Out of that study among all the above a best
approach can be derived.
VII. IMPLEMENTATION PROCEDURE
The K-Means algorithm is a typical partition-based
clustering method. Given a pre-specified number K, the
algorithm partitions the data set into K disjoint subsets
which optimize the following objective function:
Here, O is a data object in cluster Ci and muiis the centroid
[21]
(mean of objects) of Ci. Thus, the objective function E
tries to minimize the sum of the squared distances of objects
from their cluster centres.

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
232
_______________________________________________________________________________________
7.1 Algorithm:
 The K-Means algorithm accepts the "number of
clusters" to group data into and the dataset to
cluster the input values.
 The K-Means algorithm then creates the ¯rst k
initial clusters from the data set
 The K-Means algorithm calculates the arithmetic
mean of each cluster formed in the data set. The
arithmetic mean is the mean of all the individual
records in the cluster.
 Next K-Means assigns each record in the dataset to
only one of the initial clusters. Each record is
assigned to the nearest cluster using proximity
measure like Euclidean distance.
 K-Means reassigns each record in the dataset to the
most similar cluster and recalculates the arithmetic
mean of the clusters in the dataset.
 K-Means reassigns each record in the dataset to
only one of the new clusters formed.
 The preceding steps are repeated until "stable
clusters" are formed and the K-Means clustering is
completed
7.2 SELF ORGANIZING MAP (SOM)
The Self-Organizing Map (SOM) was developed by
Kohonen in 1997, on the basis of a single layered neural
network. The data objects are presented at the input, and the
output neurons are organized with a simple neighbourhood
structure such as a two dimensional p*q grid.
7.3 HIERARCHICAL CLUSTERING
Hierarchical clustering generates a hierarchical series of
nested clusters which can be graphically represented by a
tree, called dendrogram.
The branches of a dendrogram not only record the formation
of the clusters but also indicate the similarity between the
clusters.
The hierarchical clustering scheme:
Let S={Si, Sj} is the input similarity matrix, where Si,j
indicates similarity between two data objects based on
Euclidean distance.
Algorithm:
 Find a minimal entry s(i, j) in S, and merge clusters
i and j.
 Modify S by deleting rows and columns i, j and
adding a new row iand column j, with their
dissimilarities defined by:

If there is more than one cluster, then go to initial
step.
Common variants of this scheme, obtained for appropriate
choices of the parameters, are the following:
Cluster Identification Via Connectivity Kernels
(CLICK) ALGORITHMS
CLICK (Cluster Identification via Connectivity Kernels).
This utilizes graph-theoretic and statistical techniques to
identify tight groups of highly similar elements that are
likely to belong to the same true cluster. Define the
similarity between two sets as the average similarity
between their elements. An adoption step repeatedly
searches for a singleton v and a kernel K whose similarity is
maximum among all pairs of singletons and kernels. If the
value of this similarity exceeds some predefined [22]
threshold, then v is added to K and removed from R.
 K and N are the number of clusters and genes
in the data sets,
 m is a parameter which relate to `fuzziness' of
resulting clusters,
 uki is the degree of membership of gene xi in
cluster k,
 d2(xi; ck) is the distance from gene xi to
centroid ck.
The parameters in this equation are the cluster centroid
vector ck and the components of the membership
vectors uki.
These unknown parameters can be optimized by La-
grange method. Calculated uki shows the belonging
ratio to a cluster k and centroid ck shows the
representative gene expression profile of a cluster k.

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
233
_______________________________________________________________________________________
In this study, a parameter m was set to 2.0 and the
number of clusters was set to 5.
For the number of the clusters in the other clustering
method, the selected are same number as that of clusters
using fuzzy k-means clustering in order to comparing
the clustering results.
VIII. RESULTS
8.1 SCREEN SHORT FOR IMPORT GENE
EXPRESSION DATA
8.2 SCREEN SHOT FOR FINALLY PLOTTING
MEANS AND NMEANS
8.3 SCREEN SHOT FOR K-MEANS RESULT
SELF ORGANIZING MAP RESULT
8.4 GIVING THE GENE EXPRESSION DATA TO
THE CLUSTER 3.0
8.5 ANALYSIS ON GENE EXPRESSION DATA TO
THE CLUSTER 3.0
8.6 SELF ORGANIZING MAP RESULT

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
234
_______________________________________________________________________________________
8.7 SCREEN SHOT FOR SELF
ORGANIZING MAP RESULT
8.8 FORMATTED OUTPUT SCREENS
8.9 DIPICTION OF THE TREE
IX. CONCLUSION
In conclusion, cluster analysis requires experience and
knowledge about the behaviour of clustering algorithms, and
can benefit from a priori knowledge about the data and
underlying biological processes. When a priori knowledge
about the data is not available or insufficient, it may be
desirable to try different algorithms to explore the data and
get meaningful clustering results through comparisons.
However for the considered data set CLICK gives it best
with less complexity. Because of the agglorative approach
of Hierarchal Clustering too gave its best. Fuzzy is the
adaptive approach as it allows one gene to cluster among
two or three clusters.
Given the variety of available clustering algorithms, one of
the problems faced by biologists is the selection of the
algorithm most appropriate to a given gene expression data
set. A gene expression data set typically contains thousands
of genes. However, biologists oftenhave different
requirements on cluster granularity for different subsets of
genes. For some purpose, biologists may be particularly
interested in some specific subsets of genes and prefer small
and tight clusters. While for other genes, people may only
need a coarse overview of the data structure. However, most
of the existing clustering algorithms only provide a crisp set
of clusters and may not be flexible to different requirements
for cluster granularity on a single data set.
Several existing approaches, such as hierarchical clustering,
SOM, and some of the major CT algorithms can graphically
represent the cluster structure. However, these algorithms
may not be able to adapt to different userrequirements on
cluster granularity for different subsets of the data. In the
study, the functions of some genes have been studied in the
literature, which can provide guidance to the clustering.
Furthermore, some groups of the experimental conditions
are known to be strongly correlated, and the differences
among the cluster structures under these different groups
may be of particular interest. If a clustering algorithm could
integrate such partial knowledge as some clustering
constraints when carrying out the clustering task, we can
expect the clustering results would be more biologically
meaningful. In this way, clustering could cease to be a
“pure” unsupervised process and become an interactive
exploration of the data set.
References
[1] M Ramachandra & R Bramamrambha,
Implementation of classification rule discovery on
biological datsets using Ant Colony Optimization,
IJETM, 3(4): 1-12, 2016.
[2] Ben-Dor A., Shamir R. and Yakhini Z. Clustering gene
expression patterns. Journal of ComputationalBiology,
6(3/4):281–297, 1999.
[3] Arun K Pujari, Data mining Techniques, University
Press, Hyderabad, 2002.
[4] Prasad, Vadamodula, T. Srinivasa Rao, and PVGD
Prasad Reddy. "Improvised prophecy using
regularization method of machine learning algorithms on
medical data." Personalized Medicine Universe 5 (2016):
32-40.

Volume: 5 Issue: 7 228 – 235
_______________________________________________________________________________________________
235
_______________________________________________________________________________________
[5] Daxin Jiang, Chun Tang and Aidong Zhang, Cluster
Analysis for Gene expression Data: A Survey", IEEE
Transactions on Knowledge and Data Engineering, Vol.
16,No.11, November 2004, pp:1370-1386.
[6] M Ramachandra & R Bramamrambha, Comparing
clustering techniques for gene expression data analysis,
Research India Publications, 10(11): 10397-10399, 2015.
[7] Vadamodula, Prasad, et al. "Scrutiny of Data Sets
Through Procedural Algorithms for
Categorization." Data Engineering and Intelligent
Computing. Springer, Singapore, 2018. 437-444.
[8] Herrero J., Valencia A. and Dopazo J, A hierarchical
unsupervised growing neural network for clustering gene
expression patterns", Bioinformatics, Vol:17, 2001,
pp:126-136.
[9] Sharan R, Elkon R, and Shamir R, Cluster Analysis
and its Applications to GeneExpression Data", Ernest
Schering workshop on Bioinformatics and Genome
Analysis,Springer Verlog, 2002.
[10] Shamir R. and Sharan R. Click: A clustering algorithm
for gene expression analysis. In Proceedings of the 8th
International Conference on Intelligent Systems for
Molecular Biology (ISMB ’00). AAAIPress., 2000.
[11] Prasad, V., T. Srinivasa Rao, and M.
Purnachandrarao. "Proportional analysis of non linear
trained datasets on identified test datasets." International
Conference on recent trends and research issues in
computer science & engineering. Vol. 1. No. 1. 2015.
[12] Prasad, V., T. Srinivasa Rao, and B. Sai Ram.
"Information clustering based upon rough sets." Int. J.
Sci. Eng. Technol. Res.(IJSETR) 3 (2014): 8330-8333.
[13] Prasad, V., R. Siva Kumar, and M. Mamtha. "Plug in
generator to produce variant outputs for unique data." Int
J Res Eng Sci 2.4 (2014): 1-7.
[14] Prasad, V., T. Srinivas Rao, and M. Surendra Prasad
Babu. "Offline analysis & optimistic approach on
livestock expert advisory system." Artificial Intelligent
Systems and Machine Learning 5.12 (2013): 488.
[15] Prasad, V., and T. Srinivasa Rao. “Permissible thyroid
datasets assessment through kernel PC Algorithm and
Vapnik Chervonenkis theory for categorization and
classification." Data-Enabled Discovery and
Applications, 1.1 (2017): 1-11
[16] Eisen, Michael B., Spellman, Paul T., Brown, Patrick
O. and Botstein, David . Cluster analysis and display of
genome-wide expression patterns. Proc. Natl. Acad. Sci.
USA, 95(25):14863–14868, December1998.
[17] Prasad,Vadamodula.TamadaSrinivasaRao.Implement
ation of Regularization Method Ridge Regression on
Specific Medical Datasets."." International Journal of
Research in Computer Applications & Information
Technology 3 (2015): 25-33.
[18] Kohonen T. Self-Organization and Associative
Memory.Spring-Verlag, Berlin, 1984.
[19] Prasad, V., et al. "Comparative study of medical
datasets IETD and UCITD using statistical methods."
(2015).
[20] Prasad, V., T. Srinivasa Rao, and M. Surendra
Prasad Babu. "Thyroid disease diagnosis via hybrid
architecture composing rough data sets theory and
machine learning algorithms." Soft Computing 20.3
(2016): 1179-1189.
[21] M Ramachandra & R Bramamrambha, Particle
swarm optimization for initial centroid selection in
portioned clustering, IJAREEIE, 4(2): 175-184, 2015.
[22] Prasad, Vadamodula, Thamada Srinivasa Rao, and
Ankit Kumar Surana. "Standard cog exploration on
medicinal data." International Journal of Computer
Applications 119.10 (2015).

Clustering Approaches for Evaluation and Analysis on Formal Gene Expression Cancer Datasets

More Related Content

What's hot (18)

Similar to Clustering Approaches for Evaluation and Analysis on Formal Gene Expression Cancer Datasets (20)

More from rahulmonikasharma (20)

Recently uploaded (20)

Clustering Approaches for Evaluation and Analysis on Formal Gene Expression Cancer Datasets