A thorough recognition of the nature and duties of the genes is based upon having adequate information about the proteins. However, the proteomic projects follow a slow trend, therefore, solving the protein-related problems has become as one of the most important challenges in bio-informatics. Consequently, the presence of tools which can enhance the structural recognition, classification, and interpretation of proteins will be advantageous. Statistical methods are among the tools to help solve bio-informatics problems. These methods may be used to help predict the third structures of proteins, study proteins collectively, as well as extract new interactions among the protein collections. One of the very efficient and useful methods in the collective study of protein subsets is the cluster analysis. In the present study, the recognized protein sequences related to esophagus, stomach, and colon cancers are analyzed through PARTITIONING, non-PARTITIONING, and fuzzy clustering methods. Needleman-Wunsch global alignment algorithm was used to determine pair-wise similarities. The evaluations have shown that the clusters obtained through using the AGNES method have produced more powerful structures, yet, it can be said that the PAM clustering method, compared to other ones, has produced the best results in predicting ability of the 3D structure of the unknown protein sequences.