EVOLUTIONARY SYNTHESIS OF PATTERN RECOGNITION SYSTEMS
CONTENTS INTRODUCTION TO PATTERN RECOGNITION AND BIOINFORMATICS INTRODUCTION BASICS OF MOLECULAR BIOLOGY NUCLEIC ACIDS PROTEINS BIOINFORMATICS TASKS FOR BIOLOGICAL DATA ALIGNMENT AND COMPARISON OF DNA, RNA, AND PROTEIN SEQUENCES IDENTIFICATION OF GENES AND FUNCTIONAL SITES FROM DNA SEQUENCES PREDICTION OF PROTEIN FUNCTIONAL SITES DNA AND RNA STRUCTURE PREDICTION PROTEIN STRUCTURE PREDICTION AND CLASSIFICATION MOLECULAR DESIGN AND MOLECULAR DOCKING PHYLOGENETIC TREES FOR STUDYING EVOLUTIONARY RELATIONSHIP ANALYSIS OF MICROARRAY EXPRESSION DATA PATTERN RECOGNITION PERSPECTIVE PATTERN RECOGNITION RELEVANCE OF SOFT COMPUTING SCOPE AND ORGANIZATION OF THE BOOK REFERENCES PART I CLASSI?CATION NEURAL NETWORK TREE FOR IDENTI?CATION OF SPLICE JUNCTION AND PROTEIN CODING REGION IN DNA NEURAL NETWORK BASED TREE-STRUCTURED PATTERN CLASSIFIER SELECTION OF MULTILAYER PERCEPTRON SPLITTING AND STOPPING CRITERIA IDENTIFICATION OF SPLICE-JUNCTION IN DNA SEQUENCE DESCRIPTION OF DATA SET EXPERIMENTAL RESULTS IDENTIFICATION OF PROTEIN CODING REGION IN DNA SEQUENCE DATA AND METHOD FEATURE SET CONCLUSION AND DISCUSSION DESIGN OF STRING KERNEL TO PREDICT PROTEIN FUNCTIONAL SITES USING KERNEL-BASED CLASSI?ERS STRING KERNEL FOR PROTEIN FUNCTIONAL SITE IDENTIFICATION BIO-BASIS FUNCTION SELECTION OF BIO-BASIS STRINGS USING MUTUAL INFORMATION SELECTION OF BIO-BASIS STRINGS USING FISHER RATIO NOVEL STRING KERNEL FUNCTION ASYMMETRICITY OF BIOLOGICAL DISSIMILARITY NOVEL BIO-BASIS FUNCTION BIOLOGICAL DISSIMILARITY BASED STRING SELECTION METHOD FISHER RATIO USING BIOLOGICAL DISSIMILARITY NEAREST MEAN CLASSIFIER DEGREE OF RESEMBLANCE DETAILS OF THE ALGORITHM COMPUTATIONAL COMPLEXITY QUANTITATIVE MEASURE COMPACTNESS: ?INDEX CLUSTER SEPARABILITY: ?INDEX CLASS SEPARABILITY: INDEX SUPPORT VECTOR MACHINE ILLUSTRATIVE EXAMPLE PERFORMANCE OF DIFFERENT STRING SELECTION METHODS PERFORMANCE OF NOVEL BIO-BASIS FUNCTION PART II FEATURE SELECTION ROUGH SETS FOR SELECTION OF MOLECULAR DESCRIPTORS TO PREDICT BIOLOGICAL ACTIVITY OF MOLECULES BASICS OF ROUGH SETS ROUGH SET-BASED MOLECULAR DESCRIPTOR SELECTION ALGORITHM MAXIMUM RELEVANCE-MAXIMUM SIGNIFICANCE CRITERION GENERATION OF EQUIVALENCE CLASSES DESCRIPTION OF QSAR DATA SETS SUPPORT VECTOR REGRESSION METHOD OPTIMUM NUMBER OF EQUIVALENCE CLASSES PERFORMANCE ANALYSIS COMPARATIVE PERFORMANCE ANALYSIS F-INFORMATION MEASURES FOR SELECTION OF DISCRIMINATIVE GENES FROM MICROARRAY DATA GENE SELECTION USING F-INFORMATION MEASURES MINIMUM REDUNDANCY-MAXIMUM RELEVANCE CRITERION F-INFORMATION MEASURES FOR GENE SELECTION DISCRETIZATION GENE EXPRESSION DATA SETS CLASS PREDICTION METHODS ANALYSIS USING CLASS SEPARABILITY INDEX IDENTI?CATION OF DISEASE GENES USING GENE EXPRESSION AND PROTEIN–PROTEIN INTERACTION DATA INTEGRATED METHOD FOR IDENTIFYING DISEASE GENES GENE EXPRESSION DATA SET USED IDENTIFICATION OF DIFFERENTIALLY EXPRESSED GENES OVERLAP WITH KNOWN DISEASE-RELATED GENES PPI DATA AND SHORTEST PATH ANALYSIS COMPARATIVE PERFORMANCE ANALYSIS OF DIFFERENT METHODS ROUGH SETS FOR INSILICO IDENTI?CATION OF DIFFERENTIALLY EXPRESSED MIRNAS SELECTION OF DIFFERENTIALLY EXPRESSED MIRNAS RSMRMS ALGORITHM FUZZY DISCRETIZATION B 632+ ERROR RATE DATA SETS USED OPTIMUM VALUES OF DIFFERENT PARAMETERS IMPORTANCE OF B 632+ ERROR RATE ROLE OF FUZZY DISCRETIZATION METHOD PART III CLUSTERING GROUPING FUNCTIONALLY SIMILAR GENES FROM MICROARRAY DATA USING ROUGH–FUZZY CLUSTERING CLUSTERING ALGORITHMS AND VALIDITY INDICES DIFFERENT GENE CLUSTERING ALGORITHMS QUANTITATIVE MEASURES GROUPING FUNCTIONALLY SIMILAR GENES USING ROUGH–FUZZY C-MEANS ALGORITHM ROUGH–FUZZY C-MEANS INITIALIZATION METHOD IDENTIFICATION OF OPTIMUM PARAMETERS GENE EXPRESSION DATA SETS USED IMPORTANCE OF CORRELATION-BASED INITIALIZATION METHOD PERFORMANCE ANALYSIS OF DIFFERENT C-MEANS ALGORITHMS COMPARATIVE PERFORMANCE OF CLICK, SOM, AND RFCM EISEN PLOTS BIOLOGICAL SIGNIFICANCE ANALYSIS FUNCTIONAL CONSISTENCY OF CLUSTERING RESULT MUTUAL INFORMATION BASED SUPERVISED ATTRIBUTE CLUSTERING FOR MICROARRAY SAMPLE CLASSI?CATION CLUSTERING GENES FOR SAMPLE CLASSIFICATION GENE CLUSTERING: SUPERVISED VERSUS UNSUPERVISED CRITERIA FOR GENE SELECTION AND CLUSTERING SUPERVISED GENE CLUSTERING ALGORITHM SUPERVISED SIMILARITY MEASURE GENE CLUSTERING ALGORITHM FUNDAMENTAL PROPERTY OPTIMUM VALUE OF THRESHOLD QUALITATIVE ANALYSIS OF SUPERVISED CLUSTERS IMPORTANCE OF SUPERVISED SIMILARITY MEASURE IMPORTANCE OF AUGMENTED GENES PERFORMANCE OF COARSE AND FINER CLUSTERS POSSIBILISTIC BICLUSTERING FOR DISCOVERING VALUE-COHERENT OVERLAPPING –-BICLUSTERS BICLUSTERING AND POSSIBILISTIC CLUSTERING BASICS OF BICLUSTERING POSSIBILISTIC CLUSTERING POSSIBILISTIC BICLUSTERING ALGORITHM OBJECTIVE FUNCTION BICLUSTER MEANS CONVERGENCE CONDITION TERMINATION CONDITION SELECTION OF INITIAL BICLUSTERS QUANTITATIVE INDICES AVERAGE NUMBER OF GENES AVERAGE NUMBER OF CONDITIONS AVERAGE VOLUME AVERAGE MEAN SQUARED RESIDUE DEGREE OF OVERLAPPING ANALYSIS OF GENERATED BICLUSTERS COMPARATIVE ANALYSIS OF DIFFERENT METHODS FUZZY MEASURES AND WEIGHTED CO-OCCURRENCE MATRIX FOR SEGMENTATION OF BRAIN MR IMAGES FUZZY MEASURES AND CO-OCCURRENCE MATRIX FUZZY SET CO-OCCURRENCE MATRIX SECOND ORDER FUZZY CORRELATION SECOND ORDER FUZZY ENTROPY SECOND ORDER INDEX OF FUZZINESS 2D S-TYPE MEMBERSHIP FUNCTION THRESHOLDING ALGORITHM MODIFICATION OF CO-OCCURRENCE MATRIX MEASURE OF AMBIGUITY STRENGTH OF AMBIGUITY CHAPTER INTRODUCTION TO PATTERN RECOGNITION AND BIOINFORMATICS 1 1 INTRODUCTION WITH THE GAINING OF KNOWLEDGE IN DIFFERENT BRANCHES OF BIOLOGY SUCH AS MOLECULAR BIOLOGY, STRUCTURAL BIOLOGY, AND BIOCHEMISTRY, AND THE ADVANCEMENT OF TECHNOLOGIES LEAD TO THE GENERATION OF BIOLOGICAL DATA AT A PHENOMENAL RATE [286] THE ENORMOUS QUANTITY AND VARIETY OF INFORMATION ARE BEING PRODUCED FROM THE DATA OF THE MYRIAD OF PROJECTS THAT STUDY GENE EXPRESSION, DETERMINE THE PROTEIN STRUCTURES ENCODED BY THE GENES, AND DETAIL HOW THESE PRODUCTS INTERACT WITH ONE ANOTHER THIS DELUGE OF BIOLOGICALINFORMATIONHAS,INTURN,LEDTOANABSOLUTENEEDFORCOMPUTERIZEDDATABASES TO STORE, ORGANIZE, AND INDEX THE DATA, AND FOR SPECIALIZED TOOLS TO VIEW AND ANALYZE THE DATA HENCE, COMPUTERS HAVE BECOME INDISPENSABLE TO BIOLOGICAL RESEARCH SUCH AN APPROACH IS IDEAL DUE TO THE EASE WITH WHICH COMPUTERS CAN HANDLE LARGE QUANTITIES OF DATA AND PROBE THE COMPLEX DYNAMICS OBSERVED IN NATURE BIOINFORMATICS IS A MULTIDISCIPLINARY RESEARCH AREA THAT CONCEPTUALIZES BIOLOGY IN TERMS OF MOLECULES AND APPLIES INFORMATION TECHNIQUES TO UNDERSTAND AND ORGA- NIZE THE INFORMATION ASSOCIATED WITH THESE MOLECULES ON A LARGE SCALE IT INVOLVES THE DEVELOPMENT AND ADVANCEMENT OF ALGORITHMS USING TECHNIQUES INCLUDING PAT- TERN RECOGNITION, MACHINE LEARNING, APPLIED MATHEMATICS, STATISTICS, INFORMATICS, AND BIOLOGY TO ANALYZE THE COMPLETE COLLECTION OF DNA (THE GENOME), RNA (THE TRAN- SCRIPTOME), AND PROTEIN (THE PROTEOME) OF AN ORGANISM [275] MAJOR RESEARCH EFFORTS IN THIS ?ELD INCLUDE SEQUENCE ALIGNMENT AND ANALYSIS, GENE ?NDING, GENOME ANNOTA- TION, PROTEIN STRUCTURE ALIGNMENT AND PREDICTION, CLASSI?CATION OF PROTEINS, CLUSTERING AND DIMENSIONALITY REDUCTION OF MICROARRAY EXPRESSION DATA, PROTEIN–PROTEIN DOCKING OR INTERACTIONS, MODELING OF EVOLUTION, AND SO FORTH IN OTHER WORDS, BIOINFORMATICS CAN BE DESCRIBED AS THE DEVELOPMENT AND APPLICATION OF COMPUTATIONAL METHODS TO MAKE BIOLOGICAL DISCOVERIES THE ULTIMATE ATTEMPT OF THIS ?ELD IS TO DEVELOP NEW INSIGHTS INTO THE SCIENCE OF LIFE AS WELL AS CREATING A GLOBAL PERSPECTIVE, FROM WHICH THE UNIFYING PRINCIPLES OF BIOLOGY CAN BE DERIVED [20, 22, 209, 302, 377, 391] PATTERN RECOGNITION IS THE SCIENTI?C DISCIPLINE WHOSE GOAL IS THE CLASSI?CATION OF OBJECTS INTO A NUMBER OF CATEGORIES OR CLASSES IT IS THE SUBJECT OF RESEARCHING