FIRST INTERNATIONAL CONFERENCE ON ENGINEERING AND APPLIED SCIENCES OPTIMIZATION
Volumen: 1, Numero: 1, Páginas: PP 422-433
The entropic segmentation process applied to the analysis of DNA sequences allows dividing a heterogeneous sequence in subsequent homogeneous sequences. Via this approach, it is possible to determine the similitude and difference among the same sequences and other different ones. During this process, divergence measures based on information theory are used as contrast functions in order to differentiate the entities (i.e. DNA fragment sequences in this case). Some of the most common divergence measures include: Kullback-Leibler, Jeffrey's and Jensen-Shannon. These divergence measures are part of the Csiszár's measures family. In related researches, authors define as parameters, the sub-sequence sizes and divergence measure values in an arbitrary way (the justifications for those decisions are generally not well specified). This paper describes the development of an optimization system through an iterative process that will enable determining both the subsequence size and the optimum threshold value of the divergence measure. The main goal is to reduce the arbitrary components perceived in previous researches. During an earlier research conducted by the same authors of this paper, a manual tuning of the specified parameters was carried out. This approach was used as start point toward the implementation of the system proposed in this work. Four serotypes of the Dengue virus were selected as a case of study, considering that although the four serotypes have different sequences, at the same time they have the same proteins. This means that there is common and different information among the serotypes that could be useful to develop and validate the optimization system Meta-learner to optimize parameters in an entropic segmentation process to... | Request PDF. Available from: https://www.researchgate.net/publication/287454715_Meta-learner_to_optimize_parameters_in_an_entropic_segmentation_process_to_determine_proteins_in_viral_DNA_chains [accessed Feb 21 2018].