Research

1. Protein-DNA Interaction and structure-based transcription-factor binding site prediction

Protein-DNA interactions play a crucial role in the regulation of gene expression. Knowledge of protein-DNA interactions at the structural-level can provide insights into the mechanisms of gene regulation and can guide the design of novel therapeutic molecules. Our goal is to develop computational methods and resources for modeling protein-DNA interactions and predicting transcription factor (TF) binding sites on a genomic scale. More specifically, we are interested in structural aspects of protein-DNA interactions and aim to address two related questions: 1) given a 3-dimensional complex structure of a transcription factor and a DNA sequence, can we accurately evaluate the binding affinity and specificity between the transcription factor and candidate sequences? 2) In those cases where the protein-DNA complex structure is not available, can we build a reliable protein-DNA interaction model via predictive protein-DNA docking methods?

2. Prediction of Alternatively Spliced Protein Isoform Structures

Alternative splicing (AS) is an important cellular process that two or more different mature mRNAs are generated from one pre-mRNA and is considered as one of the major means to increase the proteome size and functional diversity. While high-throughput data such as expressed sequence tags (ESTs), full-length cDNAs as well as splicing microarrays have provided a genome-wide view of the evolution and regulation of alternative splicing, our general knowledge of the isoform protein structures is very limited. Little is known about how alternative splicing affects protein structures. Currently, fewer than 10 alternatively spliced isoforms with documented structures are in the Protein Data Bank (PDB)  though there are more than 13,000 known protein isoform sequences in human alone annotated in Swissprot. This knowledge gap represents a challenge as structures hold key information of isoform functions.