Extraction of frequent patterns and potential motifs from dna sequences with data mining method

ABSTRACTMS ThesisEXTRACTION OF FREQUENT PATTERNSAND POTENTIAL MOTIFS FROM DNA SEQUENCESWITH DATA MINING METHODUlaş Baran BALOĞLUFırat UniversityGraduate School of Natural and Applied SciencesDepartment of Computer Engineering2006, Page: 45In this thesis, data mining is applied on DNA datasets in order to extract potential motifsand frequent patterns. There are also other studies on this subject, called motif finding orextraction problem. However, many of those studies suffered from bad performance caused byusage of multiple sequence alignment method.The proposed method is a hybrid solution, which is based on top-down data mining andgenetic algorithm. There are two main motivations of this approach. First, we use geneticalgorithm to create a population of candidate motifs, then to create next generations according tousage of fitness function and genetic operators. Second, we use data mining in a top-downmanner to evaluate fitness of candidate motifs.The proposed method was tested on DNA sequences of E.coli bacterias in order to showits superior parts and applicability.Keywords: Data mining, bioinformatics, motif finding problem, frequent patterns