The SIJ Transactions on Computer Science Engineering & its Applications (CSEA), Vol. 1, No. 4, September-October 2013
AIS-PRMACA: Artificial Immune System
based Multiple Attractor Cellular
Automata for Strengthening PRMACA,
Promoter Region Identification
Pokkuluri Kiran Sree*, Inampudi Ramesh Babu** & S.S.S.N. Usha Devi Nedunuri***
*Research Scholar, Department of Computer Science & Engineering, Jawaharlal Nehru Technological University Hyderabad, INDIA.
E-Mail: profkiransree@gmail.com
**Professor, Department of Computer Science & Engineering, Acharya Nagarjuna University, INDIA.
***Assistant Professor, Department of Computer Science & Engineering, Jawaharlal Nehru Technological University Kakinada, INDIA.
Abstract—Identifying promoter region is an important problem in bioinformatics. Transcription of particular
gene of DNA is initiated by a promoter. So many techniques are employed in identifying these promoter
regions but all of them will attract an average accuracy of 84%. A new algorithm based on artificial immune
system AIS-PRMACA was proposed to strengthen and improve the accuracy of promoter prediction system
PRMACA. The proposed classifier is tested with BDGP and ENCODE data sets has produced an accuracy of
90.5% which is a considerable improvement when compared with the accuracy of conventional methods
available.
Keywords—Artificial Immune System; Cellular Automata; MACA; Promoter Region.
Abbreviations—Cellular Automata (CA); Dependency Vector (DV); Dependency String (DS); Genetic
Algorithm (GA); Multiple Attractor Cellular Automata (MACA).
I.
INTRODUCTION
P
ROMOTERS are responsible for most important
biochemical functions such as nutrient transportation,
making muscle fibers and important bio chemical
functions which occupy macro regions in human body.
Specifically, the Promoters are chains of amino acids and
DNA sequences, of which there are 20 different types,
coupled by peptide bonds [Miskimins et al., 1985]. The
structural hierarchy possessed by Promoters is typically
referred to as primary and tertiary region. Promoter Region
Predication from sequences of amino acid gives tremendous
value to biological community. This is because the higherlevel and secondary level [P. Kiran Sree et al., 2013],
[Miskimins et al., 1985] regions determine the function of the
Promoters and consequently, the insight into its function can
be inferred from that.
II.
RELATED WORKS IN PROMOTER
REGION IDENTIFICATION
Nearest neighbor techniques attempt to predict the region of a
central residue, within a segment of amino acids [Kiran Sree
& Dr. Inampudi Ramesh Babu, 2013; 2013A], based on the
ISSN: 2321 – 2381
known regions of homologous segments. Steen Knudsen al
[Bauer et al., 1988] has used statistical classifiers to identify
promoter regions. Techniques for region identification
include, but are not limited to, constraint programming
methods, statistical approaches to predict the probability of
an amino acid being in one of the structural elements, and
Bayesian network models. We also proposed an algorithm to
identify the promoter regions using CA with text clustering
with a lesser accuracy. We have also proposed an algorithm
[P. Kiran Sree et al., 2013] with true MACA with
considerable accuracy [Abagyan et al., 1997; Maji &
Chaudhuri, 2004].
III.
CELLULAR AUTOMATA
Cellular Automata (CA) is a simple model of a spatially
extended decentralized system, made up of a number of
individual components (cells). The communication among
constituent cells is limited to local interaction. Each
individual cell is in a specific state that changes over time
depending on the states of its neighbors. From the days of
Von Neumann who first proposed the model of Cellular
Automata (CA) [Reese, 2001; Debasis Mitra & Michael
Smith, 2004], to Wolfram‟s recent book „A New Kind of
© 2013 | Published by The Standard International Journals (The SIJ)
124
The SIJ Transactions on Computer Science Engineering & its Applications (CSEA), Vol. 1, No. 4, September-October 2013
Science‟, the simple and local neighborhood region of CA
has attracted researchers from diverse disciplines. It has been
subjected to rigorous mathematical and physical analysis
[Kiran Sree & Dr. Inampudi Ramesh Babu, 2008; 2010] for
past fifty years and its application has been proposed in
different branches of science - both social and physical.
Definition: CA is defined a four tipple <G, Z, N, F>
Where G -> Grid (Set of cells)
Z -> Set of possible cell states
N -> Set which describe cells neighborhoods
F -> Transition Function (Rules of automata)
The evolution process is directed by the popular Genetic
Algorithm (GA) with the underlying philosophy of survival
of the fittest gene. This GA framework can be adopted to
arrive at the desired CA rule region appropriate to model a
physical system. The goals of GA formulation are to enhance
the understanding of the ways CA performs computations and
to learn how CA may be evolved to perform a specific
computational task and to understand how evolution creates
complex global behavior in a locally interconnected system
of simple cells.
IV.
ARTIFICIAL IMMUNE SYSTEM (AIS)
LEARNING
AI learning involves the task of learning from a series of
examples. The end objective is to deal with any general types
of data, including cases where the number and type of
attributes may vary, and where additional layers of learning
are superimposed, with hierarchical structure of attributes and
classes. AIS learning aims to generate classifying expressions
simple enough to be understood easily by human. It must
mimic human reasoning sufficiently to provide insight into
the AIS process. Background knowledge may be exploited in
AIS learning development, but operation is assumed without
human intervention.
4.1. Artificial Immune System (AIS) Tree
AIS tree performs multistage hierarchical decision making.
Pattern recognition based on AIS trees was motivated by the
need to interpret images from remote sensing satellites. AIS
trees in particular and induction methods in general, are
associated with AIS learning to avoid the knowledge
acquisition bottleneck for expert systems. An overview of
work on AIS trees in pattern recognition can be found in.
4.2. AIS-PRMACA Tree Building
Input: Training set S = {S1, S2, · ·, SK}
Output: AIS-PRMACA Tree.
Partition(S, K)
Step 1: Generate a AIS-PRMACA with k number of
attractor basins.
Step 2: Distribute training set S into k attractor basins
(nodes).
Step 3: Evaluate the distribution in each attractor basin.
ISSN: 2321 – 2381
Step 4: If all the examples (S‟) of an attractor basin
(node) belong to a single class, then label the
attractor basin.
Step 5: If examples (S‟) of an attractor basin belong to
K‟ number of AIS-MACA classes, then,
Partition (S‟, K‟).
Step 6: Stop.
A high level comparative perspective on the
classification literature in pattern recognition and artificial
intelligence can be found in. Tree induction from a statistical
perspective, is reviewed in. A majority of work on AIS trees
[Kiran Sree & Dr. Inampudi Ramesh Babu, 2009] in AIS
learning is an offshoot of Breiman‟s work and Quinlan‟s
algorithm. Moret provided an overview of the work on
representing boolean functions as AIS trees and diagrams and
its application in pattern recognition. Safavin and Landgrebe
surveyed the literature on AIS tree classifiers, almost entirely
from a pattern recognition perspective.
The model is built describing a predefined set of data
classes. A sample set from the database, each member
belonging to one of the predefined classes, is used to train the
model. The training phase is also termed as supervised
learning of the classifier. Each member may have multiple
features/attributes. The classifier is trained based on a
specific feature/metric. Subsequent to training, the model
performs the task of prediction in the testing phase.
Prediction of the class of an input sample is done based on
some metric, typically distance metric.
4.3. Random Generation of Initial Population
To form the initial population, it must be ensured that each
solution randomly generated is a combination of an n-bit DS
with 2m number of attractor basins (Classifier #1) and an mbit DV (Classifier #2). The chromosomes are randomly
synthesized according to the following steps.
1. Randomly partition n into m number of integers such
that
n1 + n2 + · · · + nm = n.
2. For each ni, randomly generate a valid Dependency
Vector (DV).
3. Synthesize Dependency String (DS) through
concatenation of m number of DVs for Classifier #1.
4. Randomly synthesize an m-bit Dependency Vector
(DV) for Classifier #2.
5. Synthesize a chromosome through concatenation of
Classifier #1 and Classifier #2.
4.4. Mutation Algorithm
The mutation algorithm emulates the normal mutation
scheme. It makes some minimal change in the existing
chromosome of PP (Present Population) to form a new
chromosome for NP (Next Population). The chromosome in
the current GA formulation is mutated at a single point.
Figure 1 shows a simple mutation example.
© 2013 | Published by The Standard International Journals (The SIJ)
125
The SIJ Transactions on Computer Science Engineering & its Applications (CSEA), Vol. 1, No. 4, September-October 2013
V.
EXPERIMENTAL RESULTS
Experiments are conducted on EDCODE and BDGP data
sets. The output and result (Table 1) were represented below.
Figure 1: Simple Mutation
AIS-PRMACA Output
*******************************************************************************************************************
31958321; length 12001
Length of sequence: 12001
4 promoter/enhancer(s) are predicted
Promoter Pos:
6078 LDF: +16.297 TATA box at
6049
+5.597 TATAAAGT Pos: 5942 Score: +12.499
Promoter Pos:
1363 LDF: +5.235 TATA box at
1336
+6.514 AATAAAAG
Promoter Pos:
7068 LDF: +1.165 TATA box at
7039
+4.190 TAAAAATA
Promoter Pos:
9650 LDF: +1.051 TATA box at
9618
+4.491 GTTAAAAA
*******************************************************************************************************************
Table 1: Predictive Accuracy of Promoter Prediction
Amino Acid
Algorithm
DNA Sequence
Sequence
51%
46%
Bayesian
68%
72%
Normal ID-CA
74%
68%
Neural Networks
79%
72.5%
PRMACA
89.6%
91%
AIS-PRMACA
VI.
CONCLUSION
AIS-PRMACA was trained to identify promoter regions from
sequences of DNA as well as amino acid. Through analysis
and extensive testing/training are conducted for improving
the proposed classifier accuracy over several data sets. The
experiments results over ENCODE and BDGP datasets
indicate an average accuracy of the classifier is 90.5%. The
technique of artificial immune system was employed earlier
on strengthening protein coding regions and protein structure
prediction also, seems promising to improve the classifier
accuracy in many other problems related to bioinformatics.
[3]
[4]
[5]
[6]
[7]
[8]
REFERENCES
[1]
[2]
Miskimins, W. Keith, Michael P. Roberts, Alan McClelland, &
Frank H. Ruddle (1985), “Use of a Protein-Blotting Procedure
and a Specific DNA Probe to Identify Nuclear Proteins that
Recognize the Promoter Region of the Transferring Receptor
Gene”, Proceedings of the National Academy of Sciences 82,
No. 20, Pp. 6741–6744.
C.E. Bauer, D.A. Young & B.L. Marrs (1988), “Analysis of the
Rhodobacter capsulatus puf operon. Location of the OxygenRegulated Promoter Region and the Identification of an
Additional puf-encoded gene”, Journal of Biological Chemistry
263, No. 10, pp. 4820–4827.
ISSN: 2321 – 2381
[9]
[10]
R. Abagyan, S. Batalov, T. Cardozo, M. Totrov, J. Webber&
Y. Zhou (1997), “Homology Modeling with Internal
Coordinate Mechanics: Deformation Zone Mapping and
Improvements of Models via Conformational Search”,
Promoters: Region, Function and Genetics, No. 1, Pp. 29–37.
MG. Reese (2001), “Application of a Time-Delay Neural
Network to Promoter Annotation in the Drosophila
melanogaster Genome”, Comput Chem, Vol. 26, No. 1, Pp. 51–
56.
Eric E. Snyder & Gary D. Stormo (2002), “Identification of
Protein Coding Regions in Genomic DNA”, ICCS
Transactions.
Debasis Mitra & Michael Smith (2004), “Digital Sequences
Processing in Promoter Region Identification”, Innovations in
Applied Artificial Intelligence Lecture Notes in Computer
Science, Vol. 3029, Pp. 40–49.
P. Maji & P.P. Chaudhuri (2004), “FMACA: A Fuzzy Cellular
Automata based Pattern Classifier”, Proceedings of 9th
International Conference on Database Systems, Korea, Pp.
494–505.
P. Kiran Sree & Dr. Inampudi Ramesh Babu (2008), “A Novel
Promoter Coding Region Identifying Tool using Cellular
Automata Classifier with Trust-Region Method and Parallel
Scan Algorithm (NPCRITCACA)”, International Journal of
Biotechnology & Biochemistry (IJBB), Vol. 4, No. 2, Pp. 177–
189.
P. Kiran Sree & Dr. Inampudi Ramesh Babu (2009),
“Investigating an Artificial Immune System to Strengthen the
Promoter Region Identification and Promoter Coding Region
Identification
using
Cellular
Automata
Classifier”,
International Journal of Bioinformatics Research and
Applications, Vol. 5, No. 6, Pp. 647–662.
P. Kiran Sree & Dr. Inampudi Ramesh Babu (2010),
“Identification of Promoter Region in Genomic DNA using
Cellular Automata based Text Clustering”, The International
Arab Journal of Information Technology (IAJIT),Vol. 7, No. 1,
Pp. 75–78.
© 2013 | Published by The Standard International Journals (The SIJ)
126
The SIJ Transactions on Computer Science Engineering & its Applications (CSEA), Vol. 1, No. 4, September-October 2013
[11]
[12]
[13]
P. Kiran Sree, Dr. Inampudi Ramesh Babu & S.S.S.N. Usha
Devi Nedunuri (2013), “PRMACA: A Promoter Region
identification using Multiple Attractor Cellular Automata
(MACA) in ICT and Critical Infrastructure”, Proceedings of
the 48th Annual Convention of Computer Society of India, Vol.
248, Pp. 393–399.
P. Kiran Sree & Dr. Inampudi Ramesh Babu (2013),
“PSMACA: An Automated Protein Structure Prediction using
MACA (Multiple Attractor Cellular Automata)”, Journal of
Bioinformatics and Intelligent Control (JBIC), American
Scientific Publications, USA, Vol. 2, No. 3, Pp. 1–5.
P. Kiran Sree & Dr. Inampudi Ramesh Babu (2013A), “An
Extensive Report on Cellular Automata based Artificial
Immune System for Strengthening Automated Protein
Prediction”. Advances in Biomedical Engineering Research
(ABER), Science Publications (USA), Vol. 1, No. 3, Pp. 45–51.
Prof P. Kiran Sree is working as Professor
in department of CSE in BVC Engineering
College. He has published forty technical
papers in international journals and
conferences. His areas of interests include
parallel algorithms, artificial intelligence, and
compiler design and computer networks. He
was the reviewer for many IEEE Society
conferences and Journals in artificial
intelligence and networks. His bibliography is listed in Marquis
Who‟s Who in the World, 29th Edition (2012), USA.
ISSN: 2321 – 2381
Dr Inampudi Ramesh Babu is working as a
Professor in the Department of Computer
Science& Engineering, Acharya Nagarjuna
University. Also, he has been an Academic
Senate member of the same university since
2006. He held many positions in the Acharya
Nagurjuna University as Head, Director of
the Computer Centre, Chairman of the PG
Board of studies, Member of the Executive
Council, Special Officer, Additional Convener of ICET
Examinations and Convener of MCA Admissions. He is currently
supervising ten PhD students who are working in different areas of
image processing and artificial intelligence. He has published 100
papers in international journals and conferences.
Smt S. S. S. N. Usha Devi is working as
Assistant Professor in department of CSE in
University College of Engineering, JNTU
Kakinada. She has published five papers in
international
journals
and
four
in
international conferences. She is the member
of IAENG.
© 2013 | Published by The Standard International Journals (The SIJ)
127