Alexander Ku

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2018-71

May 17, 2018

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-71.pdf

Understanding binding affinities of transcription factor (TF) proteins to DNA sequence is crucial to the identification of regulatory regions that control differential gene expression across cell types. Recent advancements in ChIP-sequencing (ChIP-seq) allow us to accurately identify binding sites for a specific TF in a cellular context of interest. However, running a separate assay for each of the thousands of known TFs for a new cell type of interest is time and cost-intensive, thus motivating the need for an efficient computational method to infer experimental results of unknown experiments using prior information gathered from experiments on robustly annotated cell types. We propose an attention-based deep learning approach for learning the minimal set of epigenetic experiments required to accurately quantify transcription factor (TF) binding sites from DNA sequence.

Advisors: Joseph Gonzalez

---

BibTeX citation:

@mastersthesis{Ku:EECS-2018-71,
    Author= {Ku, Alexander},
    Title= {Epigenetic Imputation},
    School= {EECS Department, University of California, Berkeley},
    Year= {2018},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-71.html},
    Number= {UCB/EECS-2018-71},
    Abstract= {Understanding binding affinities of transcription factor (TF) proteins to DNA sequence is crucial to the identification of regulatory regions that control differential gene expression across cell types. Recent advancements in ChIP-sequencing (ChIP-seq) allow us to accurately identify binding sites for a specific TF in a cellular context of interest. However, running a separate assay for each of the thousands of known TFs for a new cell type of interest is time and cost-intensive, thus motivating the need for an efficient computational method to infer experimental results of unknown experiments using prior information gathered from experiments on robustly annotated cell types. We propose an attention-based deep learning approach for learning the minimal set of epigenetic experiments required to accurately quantify transcription factor (TF) binding sites from DNA sequence.},
}

EndNote citation:

%0 Thesis
%A Ku, Alexander 
%T Epigenetic Imputation
%I EECS Department, University of California, Berkeley
%D 2018
%8 May 17
%@ UCB/EECS-2018-71
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-71.html
%F Ku:EECS-2018-71