A toxicity data set of 58 phenols to Tetrahymena pyriformis expressed as pEC50 (Log to base 10 of EC50) was taken from literature. 70% (41 phenols) of the data was used as training set while 30% (17 phenols) was used as test set. Muti-linear Regression equations were built using the experimental pEC50as dependent variable and the various molecular descriptors as independent variables. The best Quantitative structure-toxicity relationship (QSTR) model hinted that the toxicity of phenol was dominantly influenced by octanol-water partition coefficient (XlogP) and moment of inertia (MOMI) descriptors. The results of the statistical analysis of the two parameter model include; n = 41, LOF score = 0.079, R2 = 0.6691, R2adj. = 0.6517, Q2LOO = 0.6260, F-value = 38.42. The generated QSTR model has been proven to possess statistical significance, high predictive power and wide applicability domain. Thus, it is recommended for environmental risk assessment of phenols.
