Even zooplankton can not be utilized directly by human being; it is an important food source for numerous economical fishes. Zooplankton¡¦s predator-prey interactions can affect not only global carbon fixation, but also fisheries yields directly. Copepods and ciliates are the targets of the current study, which act as critical links between classical diatom-copepod-fish webs and microbial food webs. Analyzing their predator-prey interactions can help us understand more about marine food production. The objective of this study is to investigate the differences in swimming behavior of copepods and ciliates under two environments, which are disturbances and no disturbances of predator-prey. We use five locomotive indicators (NGDR, turning rate, diffusion coefficient, kinetic energy and fractal dimension) to quantify swimming patterns. The trajectories of copepods in the undisturbed situation show circuitous, larger turning angle, and more diffusive behavior, which associate with a lower kinetic energy. The patterns of copepod movement with the presence of prey (ciliates) are contrary to the previous situation. The patterns of ciliates in the undisturbed situation are similar to those of copepods in undisturbed situation, except smaller turning angles. The trajectories of ciliates in terms of the turning and diffusive movement when predators (copepods) show up are different from those of copepods when preys (ciliates) are present. In addition to indicators, this study develops a new encoding scheme for accommodating the spatial-temporal information embedded in the original data. By analyzing the encoded data through some data mining techniques, the predator-prey interactive behaviors in the spatial scale can be easily perceived.

Bioinformatics techniques to protein secondary structure prediction mostly depend on the information available in amino acid sequence. Support vector machines (SVM) have shown strong generalization ability in a number of application areas, including protein structure prediction. In this study, a new sliding window scheme is introduced with multiple windows to form the protein data for training and testing SVM. Orthogonal encoding scheme coupled with BLOSUM62 matrix is used to make the prediction. First the prediction of binary classifiers using multiple windows is compared with single window scheme, the results shows single window not to be good in all cases. Two new classifiers are introduced for effective tertiary classification. This new classifiers use neural networks and genetic algorithms to optimize the accuracy of the tertiary classifier. The accuracy level of the new architectures are determined and compared with other studies. The tertiary architecture is better than most available techniques.

No description available.