Resources and Tools
Here is a partial list of bioinformatics tools created by our faculty members and students. If you wish to have a tool included in the list, please contact wongls@comp.nus.edu.sg. Online services will be available from our online server soon.

Tools from Kian Lee Tan's Projects

Protein Structure Comparison and Database Search

• MatAlign is a protein structure alignment method. It is more accurate that DALI and CE in a majority of test cases. It is about 2-3 times faster than DALI, and is about same speed as CE. JBCB Paper (2006) / Software / Project Homepage

• ProtDex2 is a rapid 3D protein structure retrieval system. It is much faster than TopScan and yet comparable accuracy. Bioinformatics Paper (2004) / Software / Project Homepage

• ProtClass is a method for automatic classification of 3D protein structures. It is more accurate than CPMine. It is slightly less accurate than DALI, but a lot faster. JCB Paper (2005) / Software / Project Homepage

Tools from Anthony Tung's Projects

Gene Expression Mining

• CARPENTER is a closed pattern mining algorithm that is specially catered to mine tables with large number of columns (10,000-100,000) and small number of rows (200-1000) using a concept call row enumeration or samle enumeration under the gene expression context. / KDD03 Paper / Project Homepage

• TopkIRG is an efficient rule extraction and classifier construction algorithm that find top-k covering rules for each row/sample in the gene expression in order to construct a highly accurate rule-based classifier. / SIGMOD05 Paper / Project Homepage

• RegMiner is a gene clustering algorithm that group coherent genes together even if they are shifted and scaled version of a common pattern. / ICDE06 Paper / Project Homepage

Tools from Limsoon Wong's Projects

Protein Function Prediction

• FSWeight (V2.1) is a software for inferring the function of proteins based on its indirect interaction partners. Its sensitivity and precision are superior to Neighbour Counting, Chi2, PRODISTIN, Functional Flow, and Markov Random Field approaches. It can also be used for ranking the reliability of individual protein interactions. Source Codes / Bioinformatics Paper (2006) / BMC Paper (2007) / GIW06 Paper / Project Homepage

• FSWeight (V2.2) is a software for extensive integration of diverse biological data for protein function prediction. It is more efficient and is also more sensitive and precise than a number of earlier data fusion methods for protein function prediction. It also subsumes all the features of V2.1. Source Codes / Bioinformatics Paper (2007) / Project Homepage

  Protein Complex Prediction

• Protein Complex Prediction, PCP (V1.0), is a software to predict protein complexes from protein interaction data. Its sensitivity and precision are slightly superior to RNSC and MCL; and its efficiency is considerably better. Source Codes / CSB07 Paper / Project Homepage

• Clustering based on Maximal Cliques, CMC (V2.0), is a software to predict protein complexes from protein interaction data. Its sensitivity and precision are superior to CFinder, MCL, MCode, PCP, etc. Source Codes / Bioinformatics Paper (2009) / Project Homepage

  DNA Feature Recognition

• TIS Miner is a server for recognizing protein translation initiation sites from genomic DNA sequences. Its sensitivity is about 92% and specificity is about 96%. Server / JBCB Paper (2003) / Project Homepage

• Poly(A) Signal Miner is a server for predicting poly(A) signal in human DNA sequences. Its sensitivity is about 84% and specificity is about 72%. Server / GIW03 Paper / Project Homepage

• Sirius PSB is a software package for creating accurate classifiers of functional sites in DNA/protein sequences. As a demo of the package, we have generated the Sirius PSB server for predicting protein subcellular location (~90% accuracy) and for predicting Arabidopsis polyadenylation sites (~97% sensitivity and ~97% specificity). Server / Project Homepage