| Summary: | 碩士 === 國立中山大學 === 資訊管理學系研究所 === 101 === In recent years, using microarray time series data to reconstruct gene regulatory network, has become a very popular way. However, the number of these genes are usually very large. We want to rebuild before these genes do a proper clustering, which is the each other interaction between the genes will be divided in the same group.
The way we use here is to combine multiple data sources. On the one hand avoid being affected by the impact of a single data source. When there is only one data source, data quality will have a great impact. On the other hand, we hope to have some clustering performance improvement, and improve the subsequent reconstruction of the accuracy of gene regulatory network. In our study, we combine two different types of data sources. One of source is microarray time series data, the other is the Gene Ontology. We quantify Gene Ontology, and combine with time series data. Finally, we use the partition clustering algorithm to cluster, and use Boolnet to reconstruct gene regulatory network.
After our experiment, we can obtain more great performance when we use microarray time series data and Gene Ontology simultaneously. In the following reconstruction, when the clustering result is better, we can get a better reconstruction of gene regulatory network. Therefore, our method for clustering of gene is effective and feasible.
|
|---|
