An SVM approach for the prediction of protein crystallization property using sequence derived properties

Abstract: X-ray crystallography is the most widely used method for protein 3-dimensional structure determination. Selection of target protein that can yield high quality crystal for X-ray crystallography is a challenging task. Prediction of protein crystallization propensity from sequence information is useful for the selection of target protein for crystallization. In this work, we present a machine learning method [SVMCRYS] to classify ‘amenable to crystallization’ from ‘recalcitrant to crystallization’ using sequence derived features. An overall prediction accuracy of 77.44% was obtained from 10-fold cross-validation and 81.68% from resubstitution test. Our performance comparison study shows that SVMCRYS outperforms other methods SECRET, CRYSTALP, OB-score and ParCrys with accuracy of 89.53% on TEST-RL dataset and 86.80% on TEST dataset. SVMCRYS proves to be useful in the improvement of experimental success rate by selecting correct target protein for crystallization.

Click here to download the Executables