What it does

This tool can learn the cleavage specificity of a given class of protease. In a second step this can be used to predict proteases given a cleavage site. The method assumes that the candidate cleavage point is between the two amino acids adjacent to the central position. The method is based on an efficient string kernel implemented in the Explicit Decomposition with Neighbourhood (EDeN) library. This approach uses the notion of k-mers with gaps to enumerate all possible substrings of increasing order which are used as features in an efficient linear binary classification estimator.

Example Input

>CTSL1
SSFVSNWD
>CTSL1
SSIQATTA
>CTSL1
SSLAGCQI
>CTSL1
SSLGGTVV