Our research interests lie in the area of bioinformatics. We employ computational methods to elucidate intertwined relationships between protein/gene sequences, structure, function, interactions, genome, and pathways. The ultimate goal of our research projects is to obtain new, comprehensive understanding of how structures and functions are coded in molecular sequences and how functions of molecules are orchestrated in a cell.
Specifically, we develop and apply novel computational methods for...
More information can be found on our research projects page here.
|Emap2sec is a deep learning-based tool for detecting protein secondary structures from intermediate resolution cryo-EM maps. Visit the website , the Github site, or the paper on Nature Methods.|
|MAINMAST is a de novo modeling protocol to build an entire protein 3D model directly from near-atomic resolution EM map. Visit the web site or read the paper on Nature Commn.|
|3D-SURFER 2.0 is a web platform for real-time protein surface comparison of a given protein structure against the entire PDB using 3D Zernike descriptors. It can compare the protein surface of a single chain, domain, or complex against databases of protein chains, domains, complexes, or a combination of all three, in the latest PDB. The databases are weekly updated. Visit the server or read the paper.|
|EM-SURFER is a web platform for real-time electron microscopy database search. It compares isosurface shape of a query EM map against maps in the latest EMDB. The databases are weekly updated. Visit the server or read the paper|
|PL-PatchSurfer2 is a protein-ligand virtual screening program. The program searches a binding candidate ligand of a target protein by comparing local surface patches. The binary file can be downloaded here.|
|The protein docking suite developed includes programs to perform protein-protein docking prediction, multiple protein docking, as well as protein docking prediction using predicted protein-protein interfaces. The Linux binaries can be downloaded here|
|EESG is a sequence similarity-based protein function prediction server. It employ PSI-BLAST iteratively and essentially selects GO term annotations that appear consistently in the searches. Visit the server to submit a sequence or read the paper. ESG was among top in the 1st CAFA function prediction assessment.|
|PFP is a sequence similarity-based protein function prediction server designed to predict GO annotations for a query sequence beyond what can be found by conventional database search such as BLAST.It takes into account weakly similar sequences as well as GO term associations observed in known annotations.
Visit the server to submit a sequence or read the paper. PFP had the highest total score in a function prediciton contest held at AFP-SIG'05 (ISMB2005), and also was the best in the function prediciton category at CASP7.