Servers and Softwares


RaptorX: a Web Portal for Protein Structure and Function Prediction

RaptorX web portal for protein structure and function prediction is excelling at secondary, tertiary and contact prediction for protein sequences without close homologs in the Protein Data Bank (PDB). Given a protein sequence, RaptorX predicts its secondary and tertiary structures as well as contact map, solvent accessibility, disordered regions and binding sites.
RaptorX assigns the following confidence scores to indicate the quality of a predicted 3D model: P-value for the relative global quality, GDT (global distance test) and uGDT (un-normalized GDT) for the absolute global quality, and RMSD for the absolute local quality of each residue in the model. RaptorX-Binding predicts the binding sites of a protein sequence, based upon the predicted 3D model by RaptorX.

  • RaptorX Structure Prediction predict tertiary structure of a protein sequence using templates. This shall be the main server for tertiary structure prediction. For only secondary structure, please use the Property Prediction server, which is much faster.

  • RaptorX Property Prediction predict structure properties of a protein sequence without using templates, including 3-/8-state secondary structure, solvent accessibility, and disordered regions. For tertiary structure, please use the Structure Predicton server.

  • RaptorX Contact Prediction predict contact map of a protein sequence and tertiary structure guided by predicted contacts, but without using any templates.

Lynx: an Integrated Systems Biology Platform

Lynx is a web-based database and a knowledge extraction engine. It supports annotation and analysis of high-throughput experimental data and generation of weighted hypotheses regarding genes and molecular mechanisms contributing to human phenotypes or conditions of interest. Lynx has Its underlying knowledge base (LynxKB) that integrates various classes of information from >35 public databases and private collections, as well as manually curated data from our group and collaborators.



DeepCNF-SS is a protein secondary structure prediction tool based on Deep Convolutional Neural Fields (DeepCNF). DeepCNF is a Deep Learning extension of Conditional Neural Fields (CNF), which is an integration of Conditional Random Fields (CRF) and shallow neural networks. DeepCNF can model not only complex sequence-structure relationship by a deep hierarchical architecture, but also interdependency between adjacent SS labels, so it is much more powerful than CNF.


DeepCNF-D is a protein disorder region prediction tool based on weighted Deep Convolutional Neural Fields (DeepCNF). Order/disorder prediction task is a well-known label imbalance issue, as only  6% of residues are disordered. To solve this label imbalance issue, the reciprocal of the occurrence frequencies were used for weighting during DeepCNF training and prediction.


AcconPred is a software package that helps predicting solvent accessibility and contact number of a protein simultaneously. The method is based on a shared weight multi-task learning framework under the Conditional Neural Field (CNF) model. The multi-task learning framework on a collection of related tasks provides more accurate prediction than the framework trained only on a single task.


CNFpred is a protein structure prediction program developed by our group, which achieves much more accurate sequence-template alignment by employing a probabilistic graphical model called Conditional Neural Field (CNF). CNFpred aligns one protein sequence to its remote template using a non-linear scoring function based context-specific potential that integrates both local and global information.


MRFalign is a protein homology detection program developed by our group, which distinguishes itself with other homology detection tools by representing protein families as Markov Random Fields (MRF) and align two protein families by aligning two MRFs. By using MRFs, the long-range residue interaction pattern can be incorporated into remote homology detection. An ADMM-based method is applied so that the optimization can be done efficiently.


DeepAlign is a program for pairwise protein structure alignment. Different from many other structure alignment tools, DeepAlign aligns two protein structures using evolutionary information in addition to geometric similarity. Therefore, DeepAlign can align many more evolutionarily-related residues together than other tools. That is, the alignment generated by DeepAlign makes more biological sense.


3DCOMB is a program for multiple protein structure alignment. Different from many other multiple structure alignment tools, 3DCOMB can align multiple structures with not only as many conserved cores as possible, but also high-quality pairwise alignments. 3DCOMB is unique in that it makes use of both local and global structure environments to accurately identify highly similar fragment blocks (HSFBs) among all proteins to be aligned.


PDBtool is an integrated software for handling PDB file. This tool could parse the coordinate, do SEQRES-ATOM mapping, calculate secondary structure and solvent accessibility, compute conformational letter, and reconstruct missing residues.