Search

Byline: Illhoi Yoo, Xiaohua Hu, Il-Yeol Song Document clustering has been used for better document retrieval and text mining. In this paper, we investigate if a biomedical ontology improves biomedical literature clustering performance in terms of the effectiveness and the scalability. For this investigation, we perform a comprehensive comparison study of various document clustering approaches such as hierarchical clustering methods, Bisecting K-means, K-means and Suffix Tree Clustering (STC). According to our experiment results, a biomedical ontology significantly enhances clustering quality on biomedical documents. In addition, our results show that decent document clustering approaches, such as Bisecting K-means, K-means and STC, gains some benefit from the ontology while hierarchical algorithms showing the poorest clustering quality do not reap the benefit of the biomedical ontology.

Byline: Aiguo Du, Xiaohua Hu, Yi Pan Disulphide bonds link distant portions of protein chains and provide strong structural constraints in the form of long-range interactions. Prediction and knowledge of disulphide bond connectivity is important in reducing the search space of protein conformation. In this research, we present an effective way to predict disulphide bridges by Support Vector Machine (SVM). The SVM encoding was based on experimental results on the binding motifs of protein disulphide isomerases. The physical-chemical characteristics of the flanking sequences and the linear distance between the concerned cysteine pairs were also included in the encoding. An overall pair wise accuracy of 92% was obtained for the SP39 dataset.

Byline: Xiaohua Hu Generating high quality gene clusters and identifying the underlying biological mechanism of the gene clusters are the important goals of clustering gene expression analysis. Based on this consideration, we design and develop a unified system Gene Expression Miner (GE-Miner) by integrating cluster ensemble, text clustering and multidocument summarisation and provide an environment for comprehensive gene expression data analysis. Experimental results demonstrate that our systems can obtain high quality clusters and provide concise and informative textual summary for the gene clusters.

Byline: Xiaohua Hu, Illhoi Yoo, Xiaodan Zhang, Payal Nanavati, Debjit Das This paper introduces a wavelet transformation and a cluster ensemble framework using graph theory for clustering gene expression data sets. The experiment results indicate that wavelet transformation and cluster ensemble approaches together yield better clustering results than the single best clustering algorithm on both synthetic and yeast gene expression data sets.

Byline: Xiaohua Hu Proteinoprotein interaction network is essential to understand the fundamental processes that govern cell biology. In this paper, we integrate information extraction and data mining techniques to extract and mine the scale-free proteinoprotein interaction network from biomedical literature. The experiments on around 1,600 chromatin proteins indicate that our system is very promising for mining and analysing proteinoprotein interaction network.