Biological Sequence Data Mining

摘要: Biologists have determined that the control and regulation of gene expression is primarily by relatively short sequences in region surrounding a gene. These vary length, position, redundancy, orientation, bases. Finding these fundamental problem molecular biology with important applications. Though there exist many different approaches to signal/motif (i.e. sequence) finding, 2000 Pevzner Sze reported most current motif finding algorithms are incapable detecting target signals their so-called Challenge Problem. In this paper, we show using an iterative-restart design, our new algorithm can correctly find targets. Furthermore, taking into account fact some transcription factors form dimer or even more complex structures, process sometimes involve multiple factors, extend original challenging one. We address issue combinatorial gaps variable lengths. To demonstrate efficacy algorithm, tested it on series challenge problems, compared representative motif-finding algorithms. addition, verify its feasibility real-world applications, also several regulatory families yeast genes known motifs. The purpose paper two-fold. One introduce improved biological data mining capable dealing DNA sequences. other propose research direction for general KDD community.