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

Integrative analysis of multidimensi...

Zhao, Qing.

FindBook

Google Book

Amazon

博客來

Integrative analysis of multidimensional cancer genomic data.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Integrative analysis of multidimensional cancer genomic data./
作者:	Zhao, Qing.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2015,
面頁冊數:	145 p.
附註:	Source: Dissertation Abstracts International, Volume: 77-06(E), Section: B.
Contained By:	Dissertation Abstracts International77-06B(E).
標題:	Biostatistics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10012360
ISBN:	9781339475158

Integrative analysis of multidimensional cancer genomic data.
Zhao, Qing.

Integrative analysis of multidimensional cancer genomic data. - Ann Arbor : ProQuest Dissertations & Theses, 2015 - 145 p.

Source: Dissertation Abstracts International, Volume: 77-06(E), Section: B.

Thesis (Ph.D.)--Yale University, 2015.

This item is not available from ProQuest Dissertations & Theses.

This era of cancer research has witnessed an explosion of large datasets, for example DNA sequencing data and various types of epigenetic data on disease mechanisms, such as DNA methylation data. It is essential to develop effective analytic approaches to integrate information from different genomic measurements. Accumulating evidence has shown that jointly analyzing multiple related datasets can be informative. In this dissertation, we develop novel methodologies and computational algorithms for the integrative analysis of multi-/high-dimensional datasets with the goal of fully exploiting the complex molecular architecture of cancer, underlying the etiology of cancer development, informing prognosis, and driving cancer therapeutics. We propose three different integrative analysis strategies and demonstrate the proposed methodology with both simulation studies and real data analyses.

ISBN: 9781339475158Subjects--Topical Terms:

1002712
Biostatistics.

Integrative analysis of multidimensional cancer genomic data.
LDR:05362nmm a2200361 4500 001 2128427
005 20180104132946.5
008 180830s2015 ||||||||||||||||| ||eng d
020 $a 9781339475158
035 $a (MiAaPQ)AAI10012360
035 $a AAI10012360
040 $a MiAaPQ $c MiAaPQ
100 1 $a Zhao, Qing. $3 3185824
245 1 0 $a Integrative analysis of multidimensional cancer genomic data.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2015
300 $a 145 p.
500 $a Source: Dissertation Abstracts International, Volume: 77-06(E), Section: B.
500 $a Adviser: Shuangge Ma.
502 $a Thesis (Ph.D.)--Yale University, 2015.
506 $a This item is not available from ProQuest Dissertations & Theses.
520 $a This era of cancer research has witnessed an explosion of large datasets, for example DNA sequencing data and various types of epigenetic data on disease mechanisms, such as DNA methylation data. It is essential to develop effective analytic approaches to integrate information from different genomic measurements. Accumulating evidence has shown that jointly analyzing multiple related datasets can be informative. In this dissertation, we develop novel methodologies and computational algorithms for the integrative analysis of multi-/high-dimensional datasets with the goal of fully exploiting the complex molecular architecture of cancer, underlying the etiology of cancer development, informing prognosis, and driving cancer therapeutics. We propose three different integrative analysis strategies and demonstrate the proposed methodology with both simulation studies and real data analyses.
520 $a We first exploit clinical data and four types of genomic measurement using the Additive modeling method. The analysis goal is to jointly model the multidimensional genomic measurements and calibrate their predictive power when associating them with cancer outcomes. To accommodate the high dimensionality, we extract the features using Principal Component Analysis (PCA), Partial Least Square (PLS), and Least Absolute Shrinkage and Selection Operator (Lasso), which are representative of dimension reduction and variable selection techniques, and fit Cox survival models with combined selected features. We calibrate the predictive power of each type of genomic measurement for the prognosis of four cancer types. The joint modeling also allows us to compare the predictive power of the features measured at the transcription level and that of others measured at the DNA/epigenetic level.
520 $a In the second part of the dissertation, we construct a more comprehensive model to incorporate the interconnections between different genomic measurements. The proposed analysis consists of a Decomposition step and an Integration step. In the first step, we describe the regulation of gene expressions by regulators such as DNA methylation and copy number aberration using "regulation links". It decomposes gene expressions and their regulators into different components, which accounts for the interplay between gene expressions and their regulators and residuals that are driven by other mechanisms. The three-component decomposition deciphers the underlying biological mechanisms and has several notable statistical advantages. In the second step, we link the three components generated in the first step with cancer outcomes. We show that the method improves the model performance and enhances the interpretability of the model.
520 $a We further consider a Conditioning modeling strategy. We propose a two-stage conditioning modeling framework that integrates regulatory and transcriptional data. We model the network-based transcriptional effects, detect significantly altered subnetworks, and identify important biomarkers to predict the survival outcomes. The modeling accounts for the "hierarchical" structures among regulatory and transcriptional measurements and shows a superior model estimation and marker selection property. The two-stage process also allows us to describe the systems-level transcriptional network and link the regulatory factors that underline the network structures. By connecting the altered transcriptional network with regulatory factors, we detect a closer functional relevance, elucidate the biological processes, and end up with a model with functional interpretation of (epi) genetic alterations on the phenotype.
520 $a To demonstrate the real application of our proposed methods, extensive data analyses are carried out on data from The Cancer Genome Atlas (TCGA). It is a joint and significant advancement that has been made to acquiring multiple types of genomic measurement. It provides unprecedented opportunities for conducting systematic integrated analyses of large-scale multidimensional data. We conduct careful data quality control and preprocessing and carry out extensive data analyses using the methods we propose. We discover some unknown and novel predictive markers for cancer prognosis, especially for cancer types such as melanoma (SKCM) and lung squamous cell carcinoma (LUSC).
590 $a School code: 0265.
650 4 $a Biostatistics. $3 1002712
650 4 $a Genetics. $3 530508
650 4 $a Epidemiology. $3 568544
690 $a 0308
690 $a 0369
690 $a 0766
710 2 $a Yale University. $3 515640
773 0 $t Dissertation Abstracts International $g 77-06B(E).
790 $a 0265
791 $a Ph.D.
792 $a 2015
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10012360