Identification of hot spots is an important issue in proteomics. Identifying hot spots using Digital Signal Processing (DSP) based methods is quite useful in newly discovered proteins as these methods do not require the structural information of proteins. In this paper, Modified Gabor Wavelet Transform (MGWT) was used to predict hot spots from primary amino acid sequence of protein. Incorporation of MGWT into Resonant Recognition Model (RRM) improves the prediction of the hot spots. The proposed method only requires tuning of MGWT to the characteristic frequency of the proteins' functional group, which is determined using RRM. This DSP-based technique is illustrated using several protein examples and the results are compared with the other recently reported digital signal analysis methods, viz. digital filtering and S-Transform based approaches. Relative procedural simplicity of this method over S-transform based approach and better prediction performance than digital filtering method are the novel features.