Research

In order to achieve our overall scientific career goal of understanding how small RNAs 

and Transcription Factors work together in living cells, our research objectives for the 

current time frame are (a) to identify the function of specific circuits by studying their 

dynamic behavior in vivo using Arabidopsis thaliana as a model organism, and (b) to 

apply the computational and experimental tools developed in order to compare the 

structure and function of TF-miRNA circuits within Arabidopsis and across plant 

species.  Our research program currently has two main project tracks in pursuing these 

goals:

 

(1) Regulatory Genomics: Comprehensive mapping of transcriptional control regions. 

The lack of precise data on the location of transcriptionally active regions in plant 

genomes currently hampers large-scale identification of functional cis-regulatory control 

elements.  To address this barrier, the experimental portion of this track focuses on 

mapping of transcriptional control regions using deep sequencing and open chromatin 

assays.  One project underway is transcription start site (TSS) identification for 

microRNAs and protein-coding genes in Arabidopsis root samples using nanoCAGE-

2000, a method we have developed in the laboratory which sequences tags attached to 

the 5' caps of gene transcripts using a relatively small amount of total RNA.  This 

method is important for promoter analysis of tissue-specific and cell-type specific 

samples.  The computational projects within this track include the analysis of TSS 

distribution shapes, and identification of enrichment regions for proximal promoter 

elements. A second phase of this project now underway is to overlay probable regions 

of open chromatin to obtain a set of high-confidence transcription factor binding site 

predictions upstream of TFs, other protein-coding genes, and miRNAs.

 

(2) Network Analysis: Identification of TF-miRNA gene circuits.  Projects within this track 

focus on circuit structure at three different levels of inquiry using computational 

methods:  network motif analysis, inference using functional genomics data, and 

analysis of conserved structures.  Our first project identifies over-represented network 

substructures (“network motifs”) among the three different types of relevant nodes (TFs, 

miRNAs, and non-TF genes) by using and modifying existing approaches to deal with 

the unique behaviors of each node type.  A second project is underway to develop an 

edge-weighted model which uses gene expression data to identify those over-

represented structures which are most likely to function within a specific plant cell type.  

This project begins by building a Bayes Net on small acyclic structures, and then 

identifies the best scoring model according to an appropriate metric.  A more involved 

goal is to extend this idea to analysis of networks containing directed cycles.