Computational screening of Transcriptional Regulatory Genes
碩士 === 長庚大學 === 資訊管理研究所 === 95 === Due to recent advances in biology technologies, a gene microarray chip can observe a large amount of expression of genes at a time. However, how to discover genes involved in genetic regulations from gene expression data of gene microarray chips by computer techniq...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2007
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/33768658917859686238 |
| Summary: | 碩士 === 長庚大學 === 資訊管理研究所 === 95 === Due to recent advances in biology technologies, a gene microarray chip can observe a large amount of expression of genes at a time. However, how to discover genes involved in genetic regulations from gene expression data of gene microarray chips by computer technique is still a challenge. Because of the complex mechanism in gene regulation network a non-linear model of genetic regulations is more appropriate to model gene regulation than the linear one. For above-mentioned reasons, this thesis takes advantage of promoter sequence information to reduce computational search space in the identification of gene regulation relationship. A backpropagation neural network(ANN) is used to build a non-linear gene regulatory relationship, and a feature selection method is employed to assist ANN to identify genes that involved in a genetic regulation.
This thesis presents experimental results about thirteen cell cycle genes of Saccharomyces cerevisiae in two different non-linear models. The results show that the proposed approach is very promising.
|
|---|