Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Genomic integrative analysis to improve fusion transcript detection, liquid association and biclustering

Liu, Shuchang (2017) Genomic integrative analysis to improve fusion transcript detection, liquid association and biclustering. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Download (10MB) | Preview


More data provide more possibilities. Growing number of genomic data provide new perspectives to understand some complex biological problems. Many algorithms for single-study have been developed, however, their results are not stable for small sample size or overwhelmed by study-specific signals. Taking the advantage of high throughput genomic data from multiple cohorts, in this dissertation, we are able to detect novel fusion transcripts, explore complex gene regulations and discovery disease subtypes within an integrative analysis framework.

In the first project, we evaluated 15 fusion transcript detection tools for paired-end RNA-seq data. Though no single method had distinguished performance over the others, several top tools were selected according to their F-measures. We further developed a fusion meta-caller algorithm by combining top methods to re-prioritize candidate fusion transcripts. The results showed that our meta-caller can successfully balance precision and recall compared to any single fusion detection tool.

In the second project, we extended liquid association to two meta-analytic frameworks (MetaLA and MetaMLA). Liquid association is the dynamic gene-gene correlation depending on the expression level of a third gene. Our MetaLA and MetaMLA provided stronger detection signals and more consistent and stable results compared to single-study analysis. When applied our method to five Yeast datasets related to environmental changes, genes in the top triplets were highly enriched in fundamental biological processes corresponding to environmental changes.

In the third project, we extended the plaid model from single-study analysis to multiple cohorts for bicluster detection. Our meta-biclustering algorithm can successfully discovery biclusters with higher Jaccard accuracy toward large noise and small sample size. We also introduced the concept of gap statistic for pruning parameter estimation. In addition, biclusters detected from five breast cancer mRNA expression cohorts can successfully select genes highly associated with many breast cancer related pathways and split samples with significantly different survival behaviors.

In conclusion, we improved the fusion transcripts detection, liquid association analysis and bicluster discovery through integrative-analysis frameworks. These results provided strong evidence of gene fusion structure variation, three-way gene regulation and disease subtype detection, and thus contribute to better understanding of complex disease mechanism ultimately.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Liu, Shuchangshl96@pitt.edushl96
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Tseng, Georgectseng@pitt.eductseng
Benos, Takisbenos@pitt.edubenos
Park, Yongseokyongpark@pitt.eduyongpark
Date: 12 May 2017
Date Type: Publication
Defense Date: 18 April 2017
Approval Date: 12 May 2017
Submission Date: 12 May 2017
Access Restriction: 1 year -- Restrict access to University of Pittsburgh for a period of 1 year.
Number of Pages: 181
Institution: University of Pittsburgh
Schools and Programs: School of Medicine > Computational Biology
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Integrative analysis, Fusion transcript, Liquid association, Biclustering, RNA-Seq, Microarray
Date Deposited: 12 May 2017 15:54
Last Modified: 12 May 2018 05:15


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item