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

Shrinkage procedures for mixed model...

Michigan State University.

Linked to FindBook

Google Book

Amazon

博客來

Shrinkage procedures for mixed model analyses of microarray experiments.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Shrinkage procedures for mixed model analyses of microarray experiments./
Author:	Xiao, Lan.
Description:	183 p.
Notes:	Adviser: Robert J. Tempelman.
Contained By:	Dissertation Abstracts International68-09B.
Subject:	Biology, Bioinformatics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3282230
ISBN:	9780549243090

Shrinkage procedures for mixed model analyses of microarray experiments.
Xiao, Lan.

Shrinkage procedures for mixed model analyses of microarray experiments. - 183 p.

Adviser: Robert J. Tempelman.

Thesis (Ph.D.)--Michigan State University, 2007.

Two color microarray systems are amongst the currently most popular functional genomics tools that have permeated animal science research. This novel technology facilitates the simultaneous profiling of the behavior of tens of thousands of genes under various experimental conditions.

ISBN: 9780549243090Subjects--Topical Terms:

1018415
Biology, Bioinformatics.

Shrinkage procedures for mixed model analyses of microarray experiments.
LDR:03296nam 2200289 a 45 001 861710
005 20100720
008 100720s2007 ||||||||||||||||| ||eng d
020 $a 9780549243090
035 $a (UMI)AAI3282230
035 $a AAI3282230
040 $a UMI $c UMI
100 1 $a Xiao, Lan. $3 1029433
245 1 0 $a Shrinkage procedures for mixed model analyses of microarray experiments.
300 $a 183 p.
500 $a Adviser: Robert J. Tempelman.
500 $a Source: Dissertation Abstracts International, Volume: 68-09, Section: B, page: 5667.
502 $a Thesis (Ph.D.)--Michigan State University, 2007.
520 $a Two color microarray systems are amongst the currently most popular functional genomics tools that have permeated animal science research. This novel technology facilitates the simultaneous profiling of the behavior of tens of thousands of genes under various experimental conditions.
520 $a Data generated by microarray experiments are typically influenced by a number of complex sources of systematic and random experimental variation. Mixed models provide a powerful means to account for multiple sources of variation in very general and efficient experimental designs. Now the number of hypotheses tests are a linear function of the number of genes, each test limited by generally few replicates per treatment condition due to the substantial costs of a microarray experiment. Although several Bayesian methods have been deemed effective for borrowing information across genes for the analysis of microarray data, there remains unresolved issues for more elaborate design structures characterized by differing levels of replication.
520 $a Two alternative Bayesian approaches to mixed model inference of microarray experiments are presented in this dissertation. These methods facilitate more reliable inferences on gene effects by borrowing information from the whole ensemble of genes on not just one but several layers of variability. A proposed empirical Bayes mixed model (EB-ANOVA) pools information on ANOVA mean squares for random and residual effects across genes, thereby improving sensitivity for detecting differential expression while providing adequate control of the false discovery rate (FDR). A second model (BAYESRATIO) was subsequently constructed to generalize the common correlation assumption for microarrays having two or more spots per gene, as currently implemented in the popular R software package LIMMA. The BAYESRATIO model was shown to have better performance on ROC curves and FDR control, where LIMMA was found to be too liberal for controlling FDR. A third chapter compares different image analysis software combined with the statistical methods we proposed in previous two chapters. The significantly different data features from different image software result in dissimilar statistical inferences. The findings from this work support the contention that the background adjustment may substantially reduce the precision and increase the variability of intensity estimation.
590 $a School code: 0128.
650 4 $a Biology, Bioinformatics. $3 1018415
690 $a 0715
710 2 $a Michigan State University. $3 676168
773 0 $t Dissertation Abstracts International $g 68-09B.
790 $a 0128
790 1 0 $a Tempelman, Robert J., $e advisor
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3282230