Power spectrum sequence analysis of rheumatic

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 02 Issue: 09 | Sep-2013, Available @ http://www.ijret.org 258
POWER SPECTRUM SEQUENCE ANALYSIS OF RHEUMATIC
ARTHRITIS (RA DISEASE USING DSP TECHNIQUE
K.B.Ramesh1
, Prabhu Shankar.K.S2
, B.P.Mallikarjunaswamy3
, E.T.Puttaiah4
1
Associate Professor, Dept. of Instrumentation Technology, R.V.College of Engineering, Bangalore, India
2
Biomedical Signal Processing and Instrumentation, I.T Dept., R.V.College of Engineering, Bangalore, India
3
Professor, Department of Computer Science and Engineering, SSIT, Tumkur, Karnataka, India
4
Professor, Dept. of Environmental science, Vice-chancellor, Gulbarga University, Karnataka, India
ramesh_k_b@hotmail.com, prabhushankar.k.s@gmail.com, drbpswamy@rediffmail.com,
profetputtaiah@rediffmail.com
Abstract
Digital Signal Processing (DSP) applications in bioinformatics have received great attention in recent years, where new effective
methods for genomic sequence analysis, such as the detection of coding regions, have been developed. Rheumatic Arthritis (RA) is a
chronic systemic inflammatory disease involving primarily the peripheral synovial joints. In this work, the software module has been
implemented using MATLAB 2009a which supports bioinformatics toolbox. The DSP techniques such as Fast Fourier Transform
(FFT) and Hamming window are incorporated in the algorithm. Quantitative analysis is performed using mean amplitude and mean
normalized frequency parameters computed from the generated power spectrum. The algorithm is tested for different normal and
abnormal DNA sequences available in National center of Biotechnology Information (NCBI) database.
Keywords: Digital signal processing (DSP), Fast Fourier Transform (FFT), Rheumatic Arthritis (RA).
-----------------------------------------------------------------------***--------------------------------------------------------------------------
1. INTRODUCTION
Bioinformatics is an interdisciplinary field that develops and
improves upon methods for storing, retrieving, organizing and
analyzing biological data. A major activity in bioinformatics is
to develop software tools to generate useful biological
knowledge. Bioinformatics tools aid in the comparison of
genetic and genomic data and more generally in the
understanding of evolutionary aspects of molecular biology.
The process of predicting genes in the field of bioinformatics
has been traditionally done and is often seen as being long and
expensive.
Genomic Signal Processing is a fundamental discipline that
brings to genomics the structural model-based analysis and
synthesis that form the basis of mathematically rigorous
engineering. Signals generated by the genome must be
processed to characterize their regulatory effects and their
relationship to changes at both the genotypic and phenotypic
levels. Application is generally directed towards tissue
classification and the discovery of signaling pathways. Because
transcriptional control is accomplished by a complex method
that interprets a variety of inputs, the developments of analytical
tools that detect multivariate influences on decision making
present in complex genetic networks are essential. To carry out
such an analysis, one needs appropriate analytical
methodologies. Authors suggested several DSP techniques and
methods to predict the protein coding regions of gene but still
there is requirement for developing effective modules for the
identification of protein coding regions in the DNA sequences
pertaining to several diseases and for RA in particular. Soft
computing methods neural network, genetic algorithms, fuzzy
logic for predicting exon regions generate superior results but in
spite of efficacy of these methods, their system’s
implementation is complex and thus cannot be used widely
laboratories. The existing DSP techniques filtering techniques,
DFT technique, modified Gabor wavelet method and a method
based on average magnitude difference function(AMDF) suffers
from one of the problems like inaccurate in finding the protein
coding regions due to addition of noise, requires more
processing time, and complexity. And also string matching
methods like dynamic programming and heuristic techniques
like BLAST was used to find exon regions but was not able
obtain the satisfactory results.
Rheumatic Arthritis (RA) is a chronic, progressive and
disabling auto-immune disease. Around 2-4% of world’s
population and 0.6% of India are suffering from RA. It is a
painful condition and can cause severe disability and ultimately
affects a person’s ability to carry out everyday tasks. The
diagnosis of early RA is very difficult. It is the goal of every
rheumatologist to try to prevent joint damage, the earlier and

__________________________________________________________________________________________
more aggressively someone with RA is treated, the better the
long term prognosis.
The disease is progressive and results in pain, stiffness, and
swelling of joints, which shows deformity (deviation of joints)
and ankylosis (stiffness at joints) in the late stages of the
disease. Recurring inflammation of the affected joints (i.e.,
arthritis) leads to a degradation of cartilage and to erosive
destructions (erosions) of the bone as shown in figure 1
compared to normal joint.
Signal processing techniques have been widely used in the last
decade in gene prediction and the genomic signal processing
(GSP) field received a consistent effort from researchers. The
objective of the proposed work is to develop a genomic
software module using MATLAB to detect and predict the
presence of RA. The generated results will assist doctors for the
characterization of disease, medical practioners and research
community for better understanding of the disease development.
Figure 1: Normal and Arthritic joint
This developed software module can be used by the doctors for
the characterization of the disease and to identify the cause for
disease so that proper diagnosis can be implied to the patients.
It is useful for the Pharmacists to discover new medicines for
rheumatic arthritis disease, for research communities to
understand the factors and genes responsible for patients with
rheumatic arthritis.
Deoxyribo Nucleic Acid (DNA) is made up of linear chains of
subunits called nucleotides. Each nucleotide consists of three
parts; a nitrogenous base, a five-carbon-atom sugar and a
phosphate group. The four possible nucleotide bases are
adenine (A), cytosine (C), guanine (G), and thymine (T).
However, in a number of viruses, where ribonucleic acid (RNA)
is used as building block of their genome, thymine is replaced
by uracil (U). In DNA, individual nucleotides are connected to
each other through sugar-phosphate bonds, forming a long one
dimensional chain with two distinct ends, the 5' end (upstream),
and the 3' end (downstream).
Therefore, this DNA chain or strand can symbolically be
represented by a character string consisting of four alphabet
letters A, C, G, and T. The DNA double-helix model (Figure
3.4), which was originally proposed by J. D. Watson and F. C.
H. Crick [14], holds the following interesting features [15]:
• Two DNA strands coiled around a central axis form a
right-handed double helix i.e., their turns are clockwise
when looking down the helical axis.
• The two strands are antiparallel i.e., each strand has
specific orientation and they run in opposite directions.
• The nitrogen bases of opposite strands are hydrogen
bonded.
• The purine of one strand is paired with pyrimidine of
the other strand i.e., A is paired to T and vice versa,
and G is linked to C and vice versa.
• The purine-pyrimidine pairs are being complement to
each other, thus, two strands of a single DNA molecule
are complementary to one another. Thus, if sequence 5'
– CATTGCCAGT – 3' occurs on one strand, the other
strand must have the sequence 5' – ACTGGCAATG –
3'
i.e.,
→ Strand one: 5' – C A T T G C C A G T – 3'
→ Strand two: 3' – G T A A C G G T C A – 5'
Figure 2: DNA and its building blocks [16]
A DNA sequence can be divided into genes and intragenic
spaces. The genes are responsible for protein synthesis. A gene
can be divided into two sub regions called the exons and introns
as shown in Figure 3.6.

__________________________________________________________________________________________
Figure 3: Various regions in a Gene.
Only the exons are involved in protein coding. The bases in the
exon region can be divided into groups of three adjacent bases.
Each triplet is called a codon. Scanning the gene from left to
right, a codon sequence can be defined by concatenation of the
codons in all the exons. Each codon instructs the cell machinery
to synthesize an amino acid. The codon sequence therefore
uniquely identifies an amino acid sequence which defines a
protein.
2. METHODOLOGY
FFT–DSP method is the simple method used to perform
comparative analysis of both the gene sequences of a person
with RA and for a non RA. The FFT technique is a DSP based
approach which is implemented using MATLAB. The tool
accepts these inputs from the user and performs the DSP
operations on it, and provides the user with the output power
spectrum.
Figure 4: Basic Block Diagram of the System
Figure 5: Functional Block diagram
The block diagram below depicts the steps involved in the
implementation of the project. The DNA sequence is
downloaded from the standard database NCBI of both normal
and RA sequences. The process is broken into four main
components.
The first component in the process is to convert symbolic DNA
sequences into numerical sequences since DSP tools can
process only numerical entities. The second step involves in the
processing of sequences based on a window, and the window
shape and length are the important parameters. The third stage,
the analysis tool, is the most important component in the
process. In this step, the period-3 component is extracted from
DNA sequences to differentiate between coding and non-coding
regions. Finally, the last step is the classification of exons and
introns based on the threshold value.
2.1 Complex Indicator Sequence
The four binary indicator sequences can also be replaced by just
one indicator sequence which is called as ‘Complex indicator
Sequence’, xcis (n). For a position i in the DNA sequence x(n),
the complex indicator sequence xcis(n) values are defined as:
If x (i) = A then xcis (i) = 1
x (i) = G then xcis(i) = -1
x (i) = T then xcis(i) = j
x (i) = C then xcis(i) = -j
In this mapping we have kept the purine (A and G) in the real
axis and pyrimidine (T and C) in the imaginary axis. For
example, if x (n) = ATGATCTGAA, then xcis (n) = [1 j -1 1 j -
j j -1 1 1].
2.2 Hamming Window
When using FFT analysis to study the frequency spectrum of
signals, there are limits on resolution between different
frequencies, and on detectability of a small signal in the
presence of large one. In effect, the process of measuring a
signal for a finite time is equivalent to multiplying the signal by

__________________________________________________________________________________________
a hamming function of unit amplitude. The hamming size of
351 is chosen for the analysis and it is optimized to minimize
the maximum (nearest) side lobe, giving it a height of about
one-fifth that of the Hann window, a raised cosine with simpler
coefficients.
With α = 0.54 β = 1 – α = 0.46 N = 351
Figure 6: Impulse Response of Hamming Window.
2.3 Fast Fourier Transform
The FFT is a complex-valued linear transformation from the
time domain to the frequency domain. An FFT computes the
DFT and produces exactly the same result as evaluating the
DFT definition directly; the only difference is that an FFT is
much faster.
It involves taking FFT of a hamming window of a defined size
in a DNA sequence, which is then slide across the whole length
of the sequence [2].
The value of S (k) is large at k=N/3 when a coding region is
present. We can calculate S (N/3) for short windows of data,
and then slid the window by one or more bases and S (N/3) is
recalculated. Thus we get a picture of how S (N/3) evolves
along the length of the DNA sequence. Optimal window length
of 300 is selected for better resolution and sharpness of the
peak.
2.4 General Algorithm
Step 1: Retrieve the DNA sequence from the FASTA format
using the Bio-Informatics tool box functions.
Step 2: Convert the DNA sequence into complex indicator
sequence suitable for the spectral analysis.
Step 3: Create a hamming window of length 351, which is slide
over whole length and s(n/3) periodicity is calculated.
Step 4: Compute the power spectral density by applying DSP
technique, Fast Fourier transform to the sequences.
Step 5: Project the PSD values on the spectrum for the
prediction of protein coding regions with some threshold co-
efficient.
Step6: Calculate the Mean amplitude and Mean normalized
Frequency of obtained each spectral plot.
Step7: Calculate the ratio of mean amplitude and mean
normalized frequency for the classification of normal and RA
sequence.
Figure 7: Flow chart of the proposed algorithm
3. RESULTS AND ANALYSIS
The software module for the detection of protein coding regions
of a normal and abnormal (RA) sequences has been developed
and tested for various cases. The snapshots and various test
cases of gene sequence analysis for protein coding regions
using DSP-FFT method for Complex indicator values are given
in the following section

__________________________________________________________________________________________
3.1 Results Obtained in Pre-Processing Process
Data extraction and Numerical mapping are the major steps in
the process before applying the DSP technique to the sequence.
DNA sequences of normal and RA status are taken from the
standard database National center for Biotechnology
Information (NCBI)[17] of Homo sapiens in the fasta format.
The Downloaded sequence from NCBI is retrieved into the
Matlab environment using the Bio informatics toolbox as shown
in the below fig.4.1.
Figure8: Display of the DNA sequence retrieved from the fasta
format into the Matlab using bioinformatics toolbox.
Figure 9: Display of the numerically mapped DNA sequence
retrieved from the fasta format into the Matlab using complex
indicator method.
3.2 ORF Finder
The ORF finder (Open Reading Frame) finder is a graphical
analysis tool which finds all open reading frames of a selectable
minimum size in a user’s sequence already in the database.
Figure 10: Snap shot of ORF Finder used to find the open
reading frames of the DNA sequences.
This tool identifies all open reading frames using the standard
or alternative genetic codes by just entering the reference no. in
the tool shown in the fig. 4.3.The deduced amino acid sequence
can be saved in various formats and searched against the
sequence database using the www.blastserver.com
3.3 Result analysis for Different Cases
Case 1
Reference No: NM_001199692.1
Length: 4943 bp
Generated Result:
Figure 11: Output Spectrum obtained for case 1

__________________________________________________________________________________________
X-axis: Position - length of the input sequence (Normalized
Frequency)
Y-axis: Projection co-efficient (amplitude values) calculated as
PSD (Power Spectral Density)
Mean Amplitude[Ma] = 2.6875
Mean Frequency[Mf] = 4.7
Ratio[R] = Ma/Mf = 0.519
Validation of Output for Case 1
Figure 12: Validation of Output for case 1
Case 2
Reference No: NM_001184976.1
Length: 3780 bp
Generated Result:
Frequency)
Case 3
Reference No: XM_004317778.1
Length: 3035 bp
Generated Result:
Frequency)

__________________________________________________________________________________________
The mean amplitude and mean normalized frequency and
calculated ratio for all the downloaded sequences with their
reference no.’s are tabulated in the below table 4.8
The computed ratio of mean amplitude to mean normalized
frequency, less than 1.0 indicates the normal sequences and
more than 1.0 for abnormal sequences is plotted in the bar plot
as shown in the above figure 4.16
Table 1: Complete analysis of all 14 sequences with mean
amplitude, normalized frequency and ratio[R]
Sequences
Mean
amplitude
[Ma]
Mean
frequency
[Mf]
Ratio
= Ma/
Mf
Norm 1 2.6875 4.7 0.51
Norm 2 3.65 3.67 0.90
Norm 3 3.575 3.978 0.89
Norm 4 2.222 4.2812 0.51
Norm 5 2.5577 4.5220 0.56
Norm 6 2.6764 5.4444 0.49
Norm 7 2.7831 5.2352 0.53
RA1 3.942 3.5863 1.10
RA 2 3.5862 2.1101 1.69
RA 3 3.8997 3.0833 1.26
RA 4 5.8101 3.5862 1.62
RA 5 6.6698 4.0886 1.63
RA 6 4.439 3.083 1.43
RA 7 4.024 3.272 1.22
Figure 17: Plotted graph for all sequences.
CONCLUSION AND FUTURE SCOPE
Conclusion
• In the proposed work, software module has been
developed using MATLAB which supports
Bioinformatics tool box and DSP technique has been
implemented and applied on the genomic data for the
detection and prediction of protein coding regions and
characterization of the disease.
• The proposed algorithm is efficient, requires less
processing time, high accuracy for available genomic
data and cost effective.
• The algorithm is tested for many different DNA
sequences of both normal and abnormal (RA) available
in National center of Biotechnology Information
(NCBI) database.
• Ratio of mean amplitude to mean normalized
frequency is computed so that, less than 1.0 indicates
the normal sequences and more than 1.0 for abnormal
sequences.
Future scope
• Further efforts can be made to improve the accuracy of
the system.
• Efficiency of the developed module can be still
improved by detecting the stage of disease.
• The proposed algorithm can be made as universal
standard and also can be used to predict the other
disease.

__________________________________________________________________________________________
REFERENCES
[1] Vaidyanathan P.P and Yoon B.J “Digital filters for gene
prediction applications,” IEEE Asilomar Conference.
Signals Syst. Computers, pp.306–310, Nov 2002.
[2] Vaidyanathan P.P and Yoon BJ “The role of signal-
processing concepts in genomics and proteomics”
Journal of the Franklin Institute, special issue on
Genomics, vol. 341, no. 53, pp. 111-135, 2004.
[3] Mahmood Akhtar “Comparison of Gene and exon
prediction techniques for detection short coding
regions,” International Journal of Information
Technology Vol. 11, No.8, 2005
[4] Sajid A.Marhon, Stefan C.Kremer “Gene prediction
based on DNA spectral analysis: A literature survey,”
Journal of computational biology, vol. 18, no. 4, 2011.
[5] M.K.Hota, V.K.Srivastava, “DSP technique for gene and
exon prediction taking Complex indicator sequence,”
IEEE Signal Processing Magazine, vol. 18, no. 4, pp. 8–
20, 2009.
[6] Lun Huang, Mohammad Al Bataineh, G. E. Atkin, ,
Siyun Wang, Wei Zhang “A Novel Gene Detection
Method Based on Period-3 Property,” 31st Annual
International Conference of the IEEE EMBS
Minneapolis, USA, pp.2-6, Sept 2009.
[7] Vikrant Tomar, Dipesh Gandhi, And Vijay Kumar
“Digital Signal Processing for Gene Prediction,” IEEE
Annual International conference, pp.435-439, 2008.
[8] M.Roy, S. Barman “Spectral analysis of DNA sequence
using Recursive Weiner Khinchine theorem- comparative
approach” IEEE International Conference On Recent
Trend In Information Technology, pp.315-318, 2011.
[9] Sajid A.Marhon, Stefan C.Kremer “Protein coding
region prediction based on the adaptive representation
method,” Canadian Conference on Electrical and
Computer Engineering, pp.415-418, 2011.
[10] Sylvain Robert Rivard, Jean-Gabriel Mailloux, Rachid
Beguenane and Hung Tien Bui “Design of high
performance parallelized gene predictors in Matlab”
BMC research notes, vol.5, no. 4, pp.183-192, 2012.
[11] Benjamin Y. M. Kwan, Jennifer Y. Y. Kwan, Hon
Keung Kwan “Spectral Techniques for classifying Short
Exon and Intron Sequences” International conference
on biomedical engineering and biotechnology, pp.527-
530, 2012.
[12] S.Barman, M.Roy, S.Biswas, S. Saha “Prediction of
cancer cell using digital signal processing,” International
journal of engineering, vol. 9, no. 3, pp.91-95, 2011.
[13] J.Setubal an J.Meidanis, “Introduction to computational
molecular biology” CA, PWS Publishing Company,
1999.
[14] J. D. Watson and F. C. H. Crick, "Molecular structure of
nucleic acids: a structure for deoxyribose nucleic acid,"
Nature, vol. 171, pp. 737–738, 1953.
[15] R. J. Reece, “Analysis of genes and genomes,” England,
John Wiley & Sons Ltd, 2004.
[16] Alberts, D. Bray, A. Johnson, J. Lewis, M. Raff, K.
Roberts, and P. Walter, “Essential cell biology,” NY,
Garland Publishing, 1998.
[17] National Centre for Biotechnology Information
(NCBI).Available: http://www.ncbi.nlm.nih.gov/.

Power spectrum sequence analysis of rheumatic

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Power spectrum sequence analysis of rheumatic (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Power spectrum sequence analysis of rheumatic