Skip to Main Content
A technique for the identification of hot-spot locations in proteins using digital filters is described. In this technique, the characteristic frequency of the protein sequence of interest is first determined from the consensus spectrum of the corresponding functional group. The sequence is then filtered by using a specialized narrowband bandpass digital filter in order to select the characteristic frequency. The energy of the filtered output reveals the hot-spot locations. Zero-phase filtering is used to eliminate the need of computing the phase response of the digital filter. The technique has a unique advantage over existing computational hot-spot location techniques in that it identifies the hot-spot locations solely from the amino-acid sequence of a protein, which is usually the only information initially available for a newly discovered protein molecule. This paper deals also with a MATLAB implementation of the technique that incorporates a user-friendly graphical interface. The technique is illustrated using several protein examples and the results obtained are compared with corresponding results based on biological methodologies in order to demonstrate the usefulness, accuracy, and reliability of the technique.