A gene selection method based on risk gene set using Bayesian model averaging
碩士 === 國立成功大學 === 工業與資訊管理學系碩博士班 === 97 === Gene selection and clustering techniques are usually applied for analyzing microarray data. However, most of them do not consider the risk genes presented in biological studies. Our proposed method will modify iterative Bayesian model averaging algorithm an...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2009
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/41858179879067672017 |
| Summary: | 碩士 === 國立成功大學 === 工業與資訊管理學系碩博士班 === 97 === Gene selection and clustering techniques are usually applied for analyzing microarray data. However, most of them do not consider the risk genes presented in biological studies. Our proposed method will modify iterative Bayesian model averaging algorithm and consider the risk gene set as prior information for gene selection. One major advantage of our method is that it considers more than one model to avoid overfitting. The whole method includes three stages. The genes highly correlated with any risk gene are removed at the first stage, and the remaining genes are divided into clusters at the second stage. At the final stage, a representative gene is chosen from each cluster to form a candidate gene set. We then apply the modified iterative Bayesian model averaging algorithm to select the genes in the candidate set that are suitable for deriving a regression model with risk genes. This method is tested on four well-known gene expression data sets for breast cancer and prostate cancer. The experimental results show that our gene selection method outperforms or has similar prediction accuracy to the methods proposed by other studies.
|
|---|