PreDOSS: A webserver for prediction of donor splice sites in eukaryotes

About PreDOSS

This server can be used for the prediction of donor splice sites, mainly in vertebrates, in eukaryotic gene. For the purpose of prediction, the sequence data are first encoded into numeric form and the encoded data are then used as input in machine learning techniques like support vector machine (SVM), neural networks (ANN) and random forest (RF) for the prediction purpose. The server has been trained with the human and it will be trained with the splice site data of other vertebrates like Bos taurus and others. The performance of the server has been compared with several existing prediction approaches and found better in terms of area under ROC curve. Further, the performance of ANN is found little better than the performances of SVM and RF.

Please Cite:

Meher, P. K., Sahu, T. K., Rao, A. R. and Wahi, S. D. (2016). A computational approach for prediction of donor splice sites with improved accuracy. Journal of Theoretical Biology, 404, 285-294.


Team: Prabina Kumar Meher, Tanmaya Kumar Sahu, A. R. Rao and S. D. Wahi; Contact:

Indian Agricultural Statistics Research Institute, (ICAR), Library Avenue, New Delhi - 110012