DNA Feature Recognition

Participants: Vladimir Bajic, Rajesh Chowdhary Chuan Hock Koh, Huiqing Liu, Limsoon Wong, Roland Yap, Fanfan Zeng

Background

Correct prediction of transcription start sites, translation initiation sites, gene splice sites, poly-A sites, and other functional sites from DNA sequences are important issues in genomic research. In this project, we investigate these prediction problems using the paradigm of ``feature generation, feature selection, and feature integration''. There are two reasons for our interest in such a paradigm. The first reason is that standard tool boxes can be identified and used for each of the 3 components. For example, any statistical significance test can be used for feature selection. Similarly, any machine learning method can be used for feature integration. The main challenge is in developing a ``standard'' tool box for feature generation suitable for DNA functional sites. The second reason is that features that are critical to the recognition of specific DNA functional sites are explicitly generated and selected in this paradigm. This explicitness is helpful in understanding the underlying biological mechanism of that DNA functional site.

Achievements

We investigated some of the key components of the ``feature generation, feature selection, and feature integration'' approach to DNA feature recognition. We identified a set of standard features that are suitable for accurate prediction of functional sites in DNA sequences and for understanding their underlying biological mechanism.
We developed DNAFSMiner, a prediction system based on the ``feature generation, feature selection, and feature integration'' paradigm above, that exceeds 90% sensitivity and precision for prediction of mammalian translation initiation sites and human poly-A sites.
We extended the ``feature generation, feature selection, and feature integration'' paradigm with a ``cascade'' step. Functional sites that do not have anchor motifs cannot be easily modeled using the previous paradigm but can be easily modeled using this extended paradigm. We developed Sirius PSB, a generic system for developing classifiers based on the ``feature generation, feature selection, feature integration, and cascading classifier'' paradigm. This powerful software allows a user to rapidly develop accurate classifiers for various types of functional sites in both DNA sequences and protein sequences.
As a demonstration of Sirius PSB, we used it to construct a recognizer for poly-A sites in Arabidopsis genomic sequences and achieved the highest reported equal error rates to date (95.5% on coding sequences, 85.7% on 5'UTR sequences, and 74.6% on intronic sequences). As a second demonstration of Sirius PSB, we also used it to predict (and subsequently experimentally verified) many novel septal pore proteins.
The research also got us interested in constructing explicit models for function-specific promoters, and to use such models to recognized co-regulated genes of specific functions. We approached this problem by first identifying individual features (TFBS) in function-specific promoters, and then constructing a Bayesian network to model the relationship of these features. We chose histone genes as our test case.

Selected Publications

DNAFSMiner

Fanfan Zeng, Roland Yap, Limsoon Wong. Using Feature Generation and Feature Selection for Accurate Prediction of Translation Initiation Sites. Proceedings of 13th International Conference on Genome Informatics, pages 192--200, Tokyo, Japan, December 2002. PPT
Huiqing Liu, Limsoon Wong. Data Mining Tools for Biological Sequences. Journal of Bioinformatics & Computational Biology, 1(1):139--168, April 2003. PDF
Huiqing Liu, Hao Han, Jinyan Li, Limsoon Wong. An in silico method for prediction of polyadenylation signals in human sequences. Proceedings of 14th International Conference on Genome Informatics, pages 84--93, Yokohama, December 2003. PDF / Server
Jinyan Li, Huiqing Liu, Limsoon Wong, Roland Yap. Techniques for Recognition of Translation Initiation Sites. The Practical Bioinformatician, chapter 4, pages 71--90, World Scientific, May 2004. PS
Huiqing Liu, Hao Han, Jinyan Li, Limsoon Wong. Using Amino Acid Patterns to Accurately Predict Translation Initiation Sites. In silico Biology, 4(3):255--269, 2004. PDF / Server
Huiqing Liu, Hao Han, Jinyan Li, Limsoon Wong. DNAFSMiner: A Web-Based Software Toolbox to Recognize Two Types of Functional Sites in DNA Sequences. Bioinformatics, 21:671--673, March 2005. PDF / Server

Sirius PSB
Chuan Hock Koh, Limsoon Wong. Recognition of Polyadenylation Sites from Arabidopsis Genomic Sequences. Proceedings of 18th International Conference on Genome Informatics (GIW), pages 73--82, Singapore, December 2007. PDF / Datasets & Source codes
Chuan Hock Koh, Sharene Lin, Gregory Jedd, Limsoon Wong. Sirius PSB: A Generic System for Analysis of Biological Sequences. Journal of Bioinformatics and Computational Biology, 7(6):973-990, December 2009. PDF / Sirius PSB v2.1 Software / Sirius PSB v2.33 Software / Sirius PSB Server
Julian Lai, Chuan Hock Koh, Monika Tjota, Laurent Pieuchot, Vignesh Raman, Karthik Balakrishna Chandrababu, Daiwen Yang, Limsoon Wong, Gregory Jedd. Intrinsically disordered proteins aggregate at fungal cell-to-cell channels and regulate intercellular connectivity. Proceedings of the National Academy of Sciences of the United States of America, 109(39):15781-15786, September 2012. PDF

Dragon Promoter Mapper
Rajesh Chowdhary, Limsoon Wong, Vladimir Bajic. Finding Functional Promoter Motifs by Computational Methods: A Word of Caution. International Journal of Bioinformatics Research and Applications, 2(3):282--288, 2006.
Rajesh Chowdhary, Sin Lam Tan, R. Ayesha Ali, Brent Boerlage, Limsoon Wong, Vladimir B Bajic. Dragon Promoter Mapper (DPM): A Bayesian Framework for Modeling Promoter Structures. Bioinformatics, 22:2310--2312, 2006.
Rajesh Chowdhary, Vladimir B. Bajic, Difeng Dong, Limsoon Wong, Jun S. Liu. Genome-wide analysis of regions similar to promoters of histone genes. BMC Systems Biology, 4(Suppl 1):S4, May 2010.

Dissertations

Huiqing Liu. Effective Use of Data Mining Technologies on Biological and Clinical Data. PhD thesis, Dept of Computer Science, National University of Singapore, Singapore, 2004.
Rajesh Chowdhary. A Bayesian System for Modeling Promoter Structure: A Case Study of Histone Promoters. PhD thesis, Dept of Computer Science, National University of Singapore, Singapore, 2007.
Koh Chuan Hock. A Generic System for Genomic Feature Recognition. Honours Year Project Report, Dept of Computer Science, National University of Singapore, Singapore, 2008.

Selected Presentations

Limsoon Wong. Using Feature Generation and Feature Selection for Accurate Prediction of Translation Initiation Sites. Invited talk at 1st KAIST-Singapore Joint Workshop on Bioinformatics & Natual Language Processing. KAIST, Daejon, South Korea. 24 February 2003.
Limsoon Wong. Accurate Recognition of Translation Initiation Sites. Invited talk at 10th Congress of the Federation of Asian and Oceanic Biochemists and Molecular Biologists, Bangalore, India, December 2003.
Limsoon Wong. Accurate Recognition of Translation Initiation Sites, Transcription Start Sites, and Polyadenylation Signals in Genomic Sequences. Invited seminar for University of Electronic Science and Technology of China, Chengdu, Sichuan, 25 May 2007.

Acknowledgements

This project is supported in part by the I²R-SOC Joint Lab on Knowledge Discovery from Clinical Data (7/03 - 6/07).

Last updated: 8/3/13, Limsoon Wong.