Prediction of cancer rescue p53 mutants in silico using Naïve Bayes learning methodology.

This research is focussed on predicting through Naïve Bayes learning, the possible p53 rescue mutants from amino-acid substitutions at the second, third and fourth site recombination that could reinstate normal p53 activity. The Naïve Bayes probability values of the amino-acid substitutions at the respective site-wise recombination were utilized to formulate the proposed Genetic Mutant Marker Extraction (GMME) technique that could unearth the hot spot cancer, strong rescue and weak rescue mutants. The p53 mutation records depicting the amino-acid substitutions obtained by yeast assays comprising of nearly 16,700 records, available at the University of California, Machine Learning Repository, were utilized as the training dataset for the GMME technique. The proposed GMME technique revealed the hot spot cancer mutants, strong rescue and weak rescue mutants leading to the detection of probable genetic markers for Cancer prediction from the surface regions 96-289 constituting the second, third and fourth site recombinations. Thus far, computational approaches have been able to predict rescue markers at region-specific mutations (96-105, 114-123, 130-156 and 223- 232) with respect to the second site recombination for three hot spot cancer mutants only viz, P152L, R158L and G245S. The GMME technique aimed at predicting possible rescue markers for p53 mutants at the second, third and fourth site recombinations revealing novel rescue markers for fourteen hot spot cancer mutants. Moreover, the GMME technique can be extended effectively to increasing number of recombinant sites that can be efficiently utilized to predict novel rescue markers.