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

Computing and Modeling Issues in Omi...

Wang, Wenjia,

Linked to FindBook

Google Book

Amazon

博客來

Computing and Modeling Issues in Omics Data Analysis /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Computing and Modeling Issues in Omics Data Analysis // Wenjia Wang.
Author:	Wang, Wenjia,
Description:	1 electronic resource (213 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
Contained By:	Dissertations Abstracts International86-11B.
Subject:	Biostatistics. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31936868
ISBN:	9798314882863

Computing and Modeling Issues in Omics Data Analysis /
Wang, Wenjia,

Computing and Modeling Issues in Omics Data Analysis /Wenjia Wang. - 1 electronic resource (213 pages)

Source: Dissertations Abstracts International, Volume: 86-11, Section: B.

The rapid advancement of high-throughput technologies has produced vast high-dimensional omics data, posing numerous challenges in statistical modeling and computing. This dissertation addresses these challenges across three areas: p-value combination testing, disease subtyping, and preprocessing of long-read transcriptome sequencing data.Part 1 presents an accurate, ultra-efficient strategy to compute p-values for the higher criticism (HC) test-a powerful tool for detecting rare and weak signals, such as causal variants in GWAS. Due to the lack of a closed-form solution, computing HC p-values has been difficult. I proposed an ultra-efficient computing strategy and the cross-entropy-based importance sampling method (ISCE) that can benchmark existing methods. The application to COVID-19 surveillance confirmed the viability of our computing strategy for large-scale inference.Part 2 introduces a multivariate outcome-guided Bayesian consensus clustering model (mgClustBCC) for disease subtyping. Traditional unsupervised clustering often overlooks clinically meaningful subtypes, and existing outcome-guided approaches use one clinical outcome. My model integrates high-dimensional omics data with multiple mixed-type outcomes to guide subtyping and feature selection through consensus scores that flexibly link omics and outcomes to a master clustering. Applied to lung disease and triple-negative breast cancer datasets, the model demonstrated improved identification of clinically relevant subtypes.Part 3 focuses on preprocessing and inference for long-read transcriptome sequencing (long-RNA-seq) data, which enables detection of isoform diversity and structural variation. I developed IFDlong, a versatile pipeline for gene and isoform annotation, novel isoform discovery, expression quantification via an EM algorithm, and fusion transcript profiling. It supports both bulk and single-cell long-RNA-seq across species and has demonstrated robustness and accuracy in multiple real datasets.Contribution to public health:These methodological developments address key bottlenecks in omics analysis, offering efficient computation, practical modeling, and high-quality data preprocessing. Together, they enhance disease mechanism discovery and biomarker identification, contributing to translational research and public health advancement.

English

ISBN: 9798314882863Subjects--Topical Terms:

1002712
Biostatistics.
Subjects--Index Terms:

Bayesian consensus clustering

Computing and Modeling Issues in Omics Data Analysis /
LDR:03932nmm a22004693i 4500 001 2433825
005 20250905091021.5
006 m o d
007 cr|nu||||||||
008 261011s2025 miu||||||m |||||||eng d
020 $a 9798314882863
035 $a (MiAaPQD)AAI31936868
035 $a AAI31936868
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Wang, Wenjia, $e author. $0 (orcid)0009-0000-6845-8809 $3 3823589
245 1 0 $a Computing and Modeling Issues in Omics Data Analysis / $c Wenjia Wang.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2025
300 $a 1 electronic resource (213 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
500 $a Includes supplementary digital materials.
500 $a Advisors: Tseng, George C. Committee members: Zhang, Liangliang; Deek, Rebecca; Weeks, Daniel E.; Yu, Guan.
502 $b Ph.D. $c University of Pittsburgh $d 2025.
520 $a The rapid advancement of high-throughput technologies has produced vast high-dimensional omics data, posing numerous challenges in statistical modeling and computing. This dissertation addresses these challenges across three areas: p-value combination testing, disease subtyping, and preprocessing of long-read transcriptome sequencing data.Part 1 presents an accurate, ultra-efficient strategy to compute p-values for the higher criticism (HC) test-a powerful tool for detecting rare and weak signals, such as causal variants in GWAS. Due to the lack of a closed-form solution, computing HC p-values has been difficult. I proposed an ultra-efficient computing strategy and the cross-entropy-based importance sampling method (ISCE) that can benchmark existing methods. The application to COVID-19 surveillance confirmed the viability of our computing strategy for large-scale inference.Part 2 introduces a multivariate outcome-guided Bayesian consensus clustering model (mgClustBCC) for disease subtyping. Traditional unsupervised clustering often overlooks clinically meaningful subtypes, and existing outcome-guided approaches use one clinical outcome. My model integrates high-dimensional omics data with multiple mixed-type outcomes to guide subtyping and feature selection through consensus scores that flexibly link omics and outcomes to a master clustering. Applied to lung disease and triple-negative breast cancer datasets, the model demonstrated improved identification of clinically relevant subtypes.Part 3 focuses on preprocessing and inference for long-read transcriptome sequencing (long-RNA-seq) data, which enables detection of isoform diversity and structural variation. I developed IFDlong, a versatile pipeline for gene and isoform annotation, novel isoform discovery, expression quantification via an EM algorithm, and fusion transcript profiling. It supports both bulk and single-cell long-RNA-seq across species and has demonstrated robustness and accuracy in multiple real datasets.Contribution to public health:These methodological developments address key bottlenecks in omics analysis, offering efficient computation, practical modeling, and high-quality data preprocessing. Together, they enhance disease mechanism discovery and biomarker identification, contributing to translational research and public health advancement.
546 $a English
590 $a School code: 0178
650 4 $a Biostatistics. $3 1002712
650 4 $a Cellular biology. $3 3172791
650 4 $a Oncology. $3 751006
650 4 $a Bioinformatics. $3 553671
650 4 $a Genetics. $3 530508
653 $a Bayesian consensus clustering
653 $a Global hypothesis testing
653 $a Long-read RNA-seq analysis
653 $a Molecular disease subtyping
653 $a p-value computation
690 $a 0308
690 $a 0379
690 $a 0992
690 $a 0369
690 $a 0715
710 2 $a University of Pittsburgh. $b Biostatistics. $e degree granting institution. $3 3823590
720 1 $a Tseng, George C. $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-11B.
790 $a 0178
791 $a Ph.D.
792 $a 2025
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31936868