Title

Prediction for human transcription start site using diversity measure with quadratic discriminant

Authors

Jun Lu^{1, 2,*} and Liaofu Luo^{1, 3}

Affiliation

¹Laboratory of Theoretical Biophysics, Faculty of Science and Technology, Inner Mongolia University, Hohhot 010021, P R China; ²Inner Mongolia University of Technology, Hohhot, 010051, P R China; ³Center for Theoretical Biology, Peking University, Beijing 100871, P R China

81 471 6552521

Email

lujun@imut.edu.cn; * Corresponding author

Article Type

Prediction Model

Date

received January 28, 2008; revised 17 March, 2008; accepted April 15, 2008; published April 28, 2008

Abstract

The accurate identification of promoter regions and transcription start sites is a challenge to the construction of human transcription regulation networks. Thus, an efficient prediction method based on theoretical formulation is necessary for this purpose. We used the method of increment diversity with quadratic discriminant analysis (IDQD) to predict transcription start sites (TSS). The method produced sensitivity and positive predictive value of more than 65% with positives to negatives ratio of 1:58. The performance evaluation using Receiver Operator Characteristics (ROC) showed an auROC (area under ROC) of greater than 96%. The evaluation by Precision Recall Curves (PRC) showed an auPRC (area under PRC) of about 26% for positives to negatives ratio of 1:679 and about 64% for positives to negatives ratio of 1:113. The results documented in this approach are either better or comparable to other known methods.

Keywords

promoter; transcription start site; increment of diversity; quadratic discriminant analysis

Citation

Lu and Luo, Bioinformation 2(7): 316-321 (2008)

Edited by

D. R. Flower

ISSN

0973-2063

Publisher

Biomedical Informatics

License

This is an Open Access article which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. This is distributed under the terms of the Creative Commons Attribution License.