東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Statistical methods in integrative a...

University of California, Berkeley.

Linked to FindBook

Google Book

Amazon

博客來

Statistical methods in integrative analysis of gene expression data with applications to biological pathways.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Statistical methods in integrative analysis of gene expression data with applications to biological pathways./
Author:	Teng, Siew-Leng.
Description:	135 p.
Notes:	Adviser: Haiyan Huang.
Contained By:	Dissertation Abstracts International69-03B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3306364
ISBN:	9780549529705

Statistical methods in integrative analysis of gene expression data with applications to biological pathways.
Teng, Siew-Leng.

Statistical methods in integrative analysis of gene expression data with applications to biological pathways. - 135 p.

Adviser: Haiyan Huang.

Thesis (Ph.D.)--University of California, Berkeley, 2007.

The use of gene expression data from biologically interrelated experiments is becoming increasingly important in pathway discovery to elucidate regulatory mechanisms in a biological process on a genomic scale. The major challenge is to develop statistical methods that provide a consistent formulation between the statistical methodology and the biology of the data to enable accurate inferences. This dissertation addresses this challenge in the two building blocks of a biological pathway---pathway genes and their functional relationships---by tackling three fundamental issues, (i) presence of experiment dependencies, (ii) complex data structure and (iii) high dimensional data.

ISBN: 9780549529705Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Statistical methods in integrative analysis of gene expression data with applications to biological pathways.
LDR:03294nam 2200289 a 45 001 852970
005 20100701
008 100701s2007 ||||||||||||||||| ||eng d
020 $a 9780549529705
035 $a (UMI)AAI3306364
035 $a AAI3306364
040 $a UMI $c UMI
100 1 $a Teng, Siew-Leng. $3 1019109
245 1 0 $a Statistical methods in integrative analysis of gene expression data with applications to biological pathways.
300 $a 135 p.
500 $a Adviser: Haiyan Huang.
500 $a Source: Dissertation Abstracts International, Volume: 69-03, Section: B, page: 1405.
502 $a Thesis (Ph.D.)--University of California, Berkeley, 2007.
520 $a The use of gene expression data from biologically interrelated experiments is becoming increasingly important in pathway discovery to elucidate regulatory mechanisms in a biological process on a genomic scale. The major challenge is to develop statistical methods that provide a consistent formulation between the statistical methodology and the biology of the data to enable accurate inferences. This dissertation addresses this challenge in the two building blocks of a biological pathway---pathway genes and their functional relationships---by tackling three fundamental issues, (i) presence of experiment dependencies, (ii) complex data structure and (iii) high dimensional data.
520 $a In the inference of functional gene relationships, we present a linear additive model with a Kronecker product covariance matrix to incorporate experiment dependencies and represent a complex data structure. Based on our model, we define a Knorm correlation as a new measure for functional gene relationships. A practical implementation is provided to estimate the Knorm correlation using an iterative procedure with gene sub-sampling, covariance shrinkage and bootstrapping techniques to stabilize the correlation estimate and reduce its variability. In real datasets, the Knorm correlation reports increased percentages of correctly inferred gene relationships that are supported by functional gene annotations.
520 $a We further extend the above framework into a strategy that progressively identifies genes in a target pathway using our defined gene inclusion criterion. Formulated as a hypothesis test in terms of the precision matrix, this criterion identifies a pathway gene if it has a significant partial correlation with at least one gene in the set of identified genes. This strategy can potentially identify genes co-regulated by a common set of genes and provide an interpretation for the indirect relationships between them. By starting out small, this strategy avoids incurring large estimation errors in the high dimensional space generated by the genome genes. The pathway relevance of the identified genes is continuously maintained by ensuring there is at least one connected path between any two genes. This strategy has identified promising genes involved in the glucosinolate pathway in a A. thaliana dataset; among them, 43% are already known to be in the pathway.
590 $a School code: 0028.
650 4 $a Biology, Biostatistics. $3 1018416
690 $a 0308
710 2 $a University of California, Berkeley. $3 687832
773 0 $t Dissertation Abstracts International $g 69-03B.
790 $a 0028
790 1 0 $a Huang, Haiyan, $e advisor
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3306364