- author = "Tetsushi Yada and Yasushi Totoki and Kiyoshi Asai and Masato Ishikawa",
- title = "Generation of Hidden Markov Model Describing Complex Motif in DNA Sequences",
- journal = "Transactions of Information Processing Society of Japan",
- year = "1999",
- volume = "40",
- number = "2",
- pages = "750--767",
- month = feb,
- keywords = "genetic algorithms, genetic programming",
- ISSN = "0387-5806",
- ISSN = "1882-7764",
-
URL = "
http://id.nii.ac.jp/1001/00012856/",
- size = "18 pages",
- abstract = "We have developed a method for the generation of hidden Markov model (HMM) representing complex motif in DNA sequences. The procedures of the method are as follows: (1) design of HMMs for elemental motifs in given DNA sequences; (2) construction of a complex motif HMM consisting of the elemental motif HMMs. Statistical analysis and genetic programming (GP) were applied to the respective procedures. At step (1), left-to-right HMMs were designed and their lengths were determined by a statistical significance. At step (2), probabilistic tree describing HMMs was defined and its structure was optimized by GP against a complex motif. Concatenation, probabilistic union, probabilistic closure, etc. were attached to nonterminal nodes. The elemental motif HMMs and an HMM for any a letter were attached to terminal nodes. In the method, the advance design of elemental motif HMMs and adoption of probabilistic tree as encoding rule of GP lead to efficient generation of complex motif HMM. It was observed that the generated HMM can detect the complex motif in uncharacterized DNA sequences with high accuracy. Further, the HMM is full of interesting suggestions of the complex motif. (author abst.)",
-
notes = "In Japanese.
Broken Sep 2018 http://sciencelinks.jp/j-east/article/199911/000019991199A0244916.php",
