资 源 简 介
We present here the S5CAL, a standalone program that calculates the strength of splicing sites by applying pre-defined species-specific SVM (support vector machine) models to input sequences. Our SVM models incorporated not only the sequence features such as GC content and sequence composition, but also PSSM (position-specific scoring matrix) and WAM (weight array matrix) scores that were previously used to describe the strength of splicing sites and distinguish true from control splicing sites. Our SVM models outperformed both PSSM- and WAM-based methods, in terms of sensitivity and specificity; and could also distinguish strong and weak splicing sites. This program is freely accessible. Additional resources including species-specific scoring matrix, consensus sequences and web logos are also available.
