Date: October 06, 2004 Time: (All day)
Event Type: Lecture
Genomic functional information regarding gene-gene interactions and gene-disease relations is valuable in biomedical research. However, such information resides mainly in the scientific literature and needs to be extracted and structured in order to be used by automatic systems. Natural language processing is increasingly being used for this purpose, although it inherently involves extraction errors. A new post-processing strategy that selects the most likely relations will be suggested for increasing the precision of natural language processing results. Then, the use of inferencing techniques for defining gene networks and gene-disease relations not explicitly mentioned in the literature will be discussed. Finally, data mining approaches for integrating extracted gene information with functional annotations from various resources and the application of statistical analysis techniques for evaluating the aggregated information will be presented. The information management and processing procedures presented can help in highlighting putative functional characteristics common tTime: o specific genes, supporting inferred gene-gene and gene-disease relations, and uncovering new biological knowledge.