The program performs the first step in locating of disulphide bridges in proteins: prediction of SS-bonding states of cysteins.
Procedure: The sequence is processed in steps.
1. Secondary structure is predicted for a query sequence.
2. Amino acid fragment as well as fragment of secondary structure in ?10 positions interval of each cysteine is compared with such fragments of training sets using prepared log-odds matrix, and the maximal score is defined for each set.
3. Scores of comparisons with profiles (weight matrices) constructed on positive (bounded) and negative examples are calculated for a given fragment.
4. Value of linear discriminant function is calculated based on 4 the most significant amino acid properties.
5. The resulting score computed as a linear combination of five scores listed above is used for the recognition.