Non-Parametric Analyses of the Regulatory Roles of LINE-1 Retrotransposons during Motor Neuron Differentiation
Name
park-hyunjinp-sm-eecs-2022-thesis.pdf
Description
Thesis PDF
Size
8.74 MB
Format
Adobe PDF
Checksum (MD5)
17810ccc72db336b32e82caedbee241e
Author(s)
Park, Hyunjin
Advisor(s)
Gifford, David K.
Date Issued
February 2022
Publisher
Massachusetts Institute of Technology
Abstract
Background: Repetitive elements make up a large portion of eukaryotic genomes, constituting two-thirds of human genome. Although their functional importance were recognized as early as 1950s, study of their functions is stagnated despite the advancements in next-generation sequencing due to difficulty in establishing the identities of specific elements involved in a biological process of interest, e.g. transcription factor (TF) binding, from functional data modalities such as ChIP-seq and ChIA-PET.
Results: First, I present a non-parametric, k-mer based method that overcomes analysis ambiguities introduced by short read multimapping and the incompleteness of reference genomes in low-complexity regions. I use this method to elucidate inferential evidence for the cell type-specific binding of transcription factors to specific L1 subfamilies from ChIP-seq datasets. Second, I applied a method named Mates of Chimera (MoC) to identify L1-derived extrachromosomal circular DNAs (eccDNAs) from Circulome-seq datasets. I characterized differential eccDNA compositions in ESC and MN cell types and found differential enrichment of transcription factor binding motifs in cell type specific eccDNAs. Third, I present inferential evidence consistent with the hypothesis that some low-complexity regions may participate in chromatin interactions with cis-regulatory sequences in a cell-type specific manner analogous to enhancer-promoter interactions.
Conclusion: The thesis elucidates a set of functional hypotheses concerning putative regulatory roles of repetitive elements, L1 elements in particular, that may be extrachromosomal. I base my hypotheses on a wide range of available data modalities including whole-genome sequencing, ChIP-seq, Circulome-seq, and ChIA-PET through non-parametric, k-mer based methods that do not rely on exact read alignment coordinates.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright MIT
Persistent DSpace Link
