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 significance and biologi...

University of Cincinnati.

Linked to FindBook

Google Book

Amazon

博客來

Statistical significance and biological relevance of microarray data clustering.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Statistical significance and biological relevance of microarray data clustering./
Author:	Guo, Junhai.
Description:	127 p.
Notes:	Adviser: Mario Medvedovic.
Contained By:	Dissertation Abstracts International69-02B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoeng/servlet/advanced?query=3302790
ISBN:	9780549487937

Statistical significance and biological relevance of microarray data clustering.
Guo, Junhai.

Statistical significance and biological relevance of microarray data clustering. - 127 p.

Adviser: Mario Medvedovic.

Thesis (Ph.D.)--University of Cincinnati, 2008.

DNA microarray is a powerful tool for monitoring expression level of large number of genes simultaneously. Various clustering methods have been applied to identifying groups of co-regulated genes and clinically relevant groups of samples with similar expression profiles. Due to the high dimensionality of the data and the complexity of clustering results, the statistical significance of such findings has been difficult to assess with traditional statistical hypothesis testing methods. Such methods rely on constructing the "null distribution" and calculating p-values. We investigated several approaches for assessing statistical significance of the results from hierarchical and model-based clustering methods based of their ability to control the false positive rate (FPR), false negative rate (FNR) and false discovery rates (FDR). We demonstrate that bootstrap confidence level derived for pairwise co-clusterings generated by the hierarchical clustering procedure exhibit certain level of control of FPR and FNR. For finite mixture model, by adding bootstrap and summarizing results into pairwise measurements, FPR and FNR can be controlled under certain simulated and real life data. Bootstrap also improves the performance of finite mixture model across all the data sets investigated here. Across all simulated and real world data studied here, only posterior pairwise probability (PPP) derived from Gaussian infinite mixture model (GIMM) offers control on FPR and FDR.

ISBN: 9780549487937Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Statistical significance and biological relevance of microarray data clustering.
LDR:02348nam 2200265 a 45 001 856653
005 20100709
008 100709s2008 ||||||||||||||||| ||eng d
020 $a 9780549487937
035 $a (UMI)AAI3302790
035 $a AAI3302790
040 $a UMI $c UMI
100 1 $a Guo, Junhai. $3 1023489
245 1 0 $a Statistical significance and biological relevance of microarray data clustering.
300 $a 127 p.
500 $a Adviser: Mario Medvedovic.
500 $a Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 0768.
502 $a Thesis (Ph.D.)--University of Cincinnati, 2008.
520 $a DNA microarray is a powerful tool for monitoring expression level of large number of genes simultaneously. Various clustering methods have been applied to identifying groups of co-regulated genes and clinically relevant groups of samples with similar expression profiles. Due to the high dimensionality of the data and the complexity of clustering results, the statistical significance of such findings has been difficult to assess with traditional statistical hypothesis testing methods. Such methods rely on constructing the "null distribution" and calculating p-values. We investigated several approaches for assessing statistical significance of the results from hierarchical and model-based clustering methods based of their ability to control the false positive rate (FPR), false negative rate (FNR) and false discovery rates (FDR). We demonstrate that bootstrap confidence level derived for pairwise co-clusterings generated by the hierarchical clustering procedure exhibit certain level of control of FPR and FNR. For finite mixture model, by adding bootstrap and summarizing results into pairwise measurements, FPR and FNR can be controlled under certain simulated and real life data. Bootstrap also improves the performance of finite mixture model across all the data sets investigated here. Across all simulated and real world data studied here, only posterior pairwise probability (PPP) derived from Gaussian infinite mixture model (GIMM) offers control on FPR and FDR.
590 $a School code: 0045.
650 4 $a Biology, Biostatistics. $3 1018416
690 $a 0308
710 2 $a University of Cincinnati. $3 960309
773 0 $t Dissertation Abstracts International $g 69-02B.
790 $a 0045
790 1 0 $a Medvedovic, Mario, $e advisor
791 $a Ph.D.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoeng/servlet/advanced?query=3302790