Browsing by Author "Chang, Hsiu-Ming"

Now showing 1 - 3 of 3

Constructing Neuronal Structure from 3D Confocal Microscopic Images
(2009)
Lee, Ping-Chang
;
Chang, Hsiu-Ming
;
Lin, Chih-Yang
;
Chiang, Ann-Shyn
;
Ching, Yu-Tai
A semiautomatic method, based on the gradient flow (GVF) snake, to construct the neuronal structure from 3D confocal microscopic images is presented. A single neuron can be labeled by using green fluorescent protein (GFP) such that the 3D neuronal structure can be visualized in a stack of confocal microscopic images. To construct the neuronal structure, we traced a target fiber by providing a rough initial path to approximate that fiber. The path is then deformed under the gradient vector field to converge to the centerline of the fiber by using the GVF snake method. Using our developed software, a neuronal sturctural can be efficiently reconstructed and the tracing results can be reduced to the tree structure of the neuron. Given the tree structure of the neuron, quantitative information can be derived and further analysis can be carried out.
Feature selection and combination criteria for improving accuracy in protein structure prediction
(2007-06-01)
Lin, Ken-Li
;
Lin, Chun-Yuan
;
Huang, Chuen-Der
;
Chang, Hsiu-Ming
;
Yang, Chiao-Yun
The classification of protein structures is essential for their function determination in bioinformatics. At present, a reasonably high rate of prediction accuracy has been achieved in classifying proteins into four classes in the SCOP database according to their primary amino acid sequences. However, for further classification into fine-grained folding categories, especially when the number of possible folding patterns as those defined in the SCOP database is large, it is still quite a challenge. In our previous work, we have proposed a two-level classification strategy called hierarchical learning architecture (HLA) using neural networks and two indirect coding features to differentiate proteins according to their classes and folding patterns, which achieved an accuracy rate of 65.5%. In this paper, we use a combinatorial fusion technique to facilitate feature selection and combination for improving predictive accuracy in protein structure classification. When applying various criteria in combinatorial fusion to the protein fold prediction approach using neural networks with HLA and the radial basis function network (RBFN), the resulting classification has an overall prediction accuracy rate of 87% for four classes and 69.6% for 27 folding categories. These rates are significantly higher than the accuracy rate of 56.5% previously obtained by Ding and Dubchak. Our results demonstrate that data fusion is a viable method for feature selection and combination in the prediction and classification of protein structure.
Three-Dimensional Reconstruction of Brain-wide Wiring Networks in Drosophila at Single-Cell Resolution
(2011-01-11)
Chiang, Ann-Shyn
;
Lin, Chih-Yung
;
Chuang, Chao-Chun
;
Chang, Hsiu-Ming
;
Hsieh, Chang-Huain
Background: Animal behavior is governed by the activity of interconnected brain circuits. Comprehensive brain wiring maps are thus needed in order to formulate hypotheses about information flow and also to guide genetic manipulations aimed at understanding how genes and circuits orchestrate complex behaviors. Results: To assemble this map, we deconstructed the adult Drosophila brain into approximately 16,000. single neurons and reconstructed them into a common standardized framework to produce a virtual fly brain. We have constructed a mesoscopic map and found that it consists of 41 local processing units (LPUs), six hubs, and 58 tracts covering the whole Drosophila brain. Despite individual local variation, the architecture of the Drosophila brain shows invariance for both the aggregation of local neurons (LNs) within specific LPUs and for the connectivity of projection neurons (PNs) between the same set of LPUs. An open-access image database, named FlyCircuit, has been constructed for online data archiving, mining, analysis, and three-dimensional visualization of all single neurons, brain-wide LPUs, their wiring diagrams, and neural tracts. Conclusion: We found that the Drosophila brain is assembled from families of multiple LPUs and their interconnections. This provides an essential first step in the analysis of information processing within and between neurons in a complete brain.