東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Machine learning for the analysis of...

Holloway, Dustin T.

FindBook

Google Book

Amazon

博客來

Machine learning for the analysis of transciptional regulation and biological systems.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Machine learning for the analysis of transciptional regulation and biological systems./
作者:	Holloway, Dustin T.
面頁冊數:	286 p.
附註:	Adviser: Charles DeLisi.
Contained By:	Dissertation Abstracts International68-04B.
標題:	Biology, Bioinformatics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3259828

Machine learning for the analysis of transciptional regulation and biological systems.
Holloway, Dustin T.

Machine learning for the analysis of transciptional regulation and biological systems. - 286 p.

Adviser: Charles DeLisi.

Thesis (Ph.D.)--Boston University, 2007.

Many factors influence the regulation of genes and their protein products within the cell. The primary mode of regulatory control is the association of transcription factors (TFs) with their binding sites in DNA. These binding sites occur most often in a gene's promoter regions. The network of interactions between transcription factors and the genes they regulate governs many of the behaviors and responses of cells. One of the central goals of modern computational biology is the ability to predict the targets of transcription factors, thereby revealing the genetic program of the cell. Combining various kinds of data (e.g., sequence, gene expression) in an optimal way to make these predictions is also a central theme of regulatory analysis. This thesis presents a data mining methodology making use of support vector machines (SVMs) and other machine learning techniques to predict new targets for transcription factors using a variety of genomic information. The employed methods allow extraction of detailed biological information from the datasets under study. Extracting this information allows us to generate hypotheses about TF function such as candidate sequences that a TF may bind or the experimental conditions under which it may act.Subjects--Topical Terms:

1018415
Biology, Bioinformatics.

Machine learning for the analysis of transciptional regulation and biological systems.
LDR:03102nam 2200265 a 45 001 947571
005 20110524
008 110524s2007 ||||||||||||||||| ||eng d
035 $a (UMI)AAI3259828
035 $a AAI3259828
040 $a UMI $c UMI
100 1 $a Holloway, Dustin T. $3 1271038
245 1 0 $a Machine learning for the analysis of transciptional regulation and biological systems.
300 $a 286 p.
500 $a Adviser: Charles DeLisi.
500 $a Source: Dissertation Abstracts International, Volume: 68-04, Section: B, page: 2008.
502 $a Thesis (Ph.D.)--Boston University, 2007.
520 $a Many factors influence the regulation of genes and their protein products within the cell. The primary mode of regulatory control is the association of transcription factors (TFs) with their binding sites in DNA. These binding sites occur most often in a gene's promoter regions. The network of interactions between transcription factors and the genes they regulate governs many of the behaviors and responses of cells. One of the central goals of modern computational biology is the ability to predict the targets of transcription factors, thereby revealing the genetic program of the cell. Combining various kinds of data (e.g., sequence, gene expression) in an optimal way to make these predictions is also a central theme of regulatory analysis. This thesis presents a data mining methodology making use of support vector machines (SVMs) and other machine learning techniques to predict new targets for transcription factors using a variety of genomic information. The employed methods allow extraction of detailed biological information from the datasets under study. Extracting this information allows us to generate hypotheses about TF function such as candidate sequences that a TF may bind or the experimental conditions under which it may act.
520 $a We first develop a Bayesian approach to combine heterogeneous data. Then, an SVM method is implemented which greatly improves on the data integration. The SVM methods are applied to 163 yeast and 153 human TFs, generating thousands of high confidence predictions. These predictions are analyzed extensively, and two case studies are presented. Specifically, new roles for the yeast regulator Swi6 are discussed along with the role of Wt1 in human cancer. Along those lines, SVMs are also applied to DNA microarrays to discover biomarkers for renal cell carcinoma which can accurately differentiate normal from cancerous tissue. In addition, several machine learning algorithms are combined to assign functions to unannotated genes in the yeast genome by integrating large genomic datasets. Finally, a new motif discovery method (SVMotif) which leverages the power of kernel methods is proposed. The work concludes with thoughts on future directions and perspectives on future extensions of this work.
590 $a School code: 0017.
650 4 $a Biology, Bioinformatics. $3 1018415
690 $a 0715
710 2 $a Boston University. $3 1017454
773 0 $t Dissertation Abstracts International $g 68-04B.
790 $a 0017
790 1 0 $a DeLisi, Charles, $e advisor
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3259828