Bioinformatic methods for analyses of epigenome and RNA interactome

The sequenced genetic codes for multiple species have provided great sources for understanding how a single genome can give rise to a complex organism. However, only a small percent of mammalian genome could be identified as protein-coding genes. Our understanding of the genome is far from complete, particularly on non-coding RNAs and regulatory sequences. Epigenomic modifications and variations were considered to contribute to the diversity of functions found across different cell types, play key roles in the establishment and maintenance of cellular identity during development. Here, we propose to annotate functional regulatory sequences based on combinatorial patterns of epigenomic marks. We will classify these patterns in two directions, through conservation information among mammalian species and through a mouse differentiation process. We are also trying to explore new functions of non-coding RNAs through analysis of genome-wide RNA-RNA interactions. The first part of my thesis is focused on the cross-species direction. We provide a comparative approach to compare epigenomic patterns in pluripotent stem cells of three different mammalian species and find that certain combinations of epigenomic modifications tend to be co-localized. These co-localizations are also more likely to be present in conserved regions. Moreover, our results suggest that these conserved co-localization patterns could help to define strong regulatory elements within the genome. The functions of these locations are tested through a guided differentiation system. The second part is about the other direction. With time-course epigenomic data from the guided differentiation system, genomic sequences are clustered based on the spatiotemporal epigenomic information. These analyses provide us better understanding how the single genome is dynamically regulated to ensure a diversity of cell types. A two-layer hierarchical model is presented and by applying this model, context-specific functions of some epigenomic modifications are uncovered. Meanwhile, a genome-wide view of regulatory sequence locations as well as their activation time is provided to help clarify the transcription network. The last part of my thesis is about a data analysis tool set and analytical results for a novel next generation sequencing based technology to identify RNA-RNA interactions. The technology allows simultaneous and unbiased identification of different types of RNA-RNA interactions within the cells. Our computational tools are able to evaluate the quality of sequencing data generated from this technology, identify strong interactions starting from raw sequencing reads and compare similarities of results from different samples. We provide two different types of visualizations for identified interactions. Novel regulatory functions of some small RNAs are also discovered in mouse ES cells. The interactions within the same RNA molecule can also provide useful information for in vivo RNA structures.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Biophysics & Computnl Biology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Yu, Pengfei

Contributors dc:contributor

• Zhong, Sheng
• Stubbs, Lisa J.
• Sinha, Saurabh
• Ma, Jian

• Bioinformatics
• comparative epigenomics
• spatiotemporal clustering
• dynamic epigenomic changes
• RNA-RNA interaction

Statement dc:rights

• Copyright 2014 Pengfei Yu

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/72957

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/72957