Background: Turns are a critical element of the structure of a protein ; turns play a crucial role in loops, folds, and interactions.Current prediction methods are well developed for the prediction of individual turn types, including α-turn, β-turn, and π-turn, etc.However, methods for analyzing and predicting turns in a protein with a uniform model have not been reported.Method: In this study, we present a novel approach, TurnP, which offers the ability to investigate all the turns in a protein for the first time.The main characteristic of TurnP is using newly exploited features of structural evolution information based on structural homologies to predict global turns in proteins.TurnP utilizes predicted secondary structures and predicted shape strings, both of which have greater accuracy, based on innovative technologies which were both developed by our group.Then, sequence and structural evolution features, which are profile of sequence, profile of predicted secondary structures and profile of predicted shape strings are generated by sequence and structure alignment.Results: When TurnP was validated on a non-redundant dataset (4,107 entries) by five-fold cross-validation, we achieved an accuracy of 88.8% and a sensitivity of 71.8%, which exceeded (was superior to) the most state-of-the-art predictors of certain type of turn.When our approach was tested with newly determined sequences (736 entries), an accuracy of 82.6% and a sensitivity of 60.8% were obtained.To test the proposed approach more rigorously, we used the EVA and CASP9 datasets as benchmarks and obtained accuracies of 81.7% and 80.2% respectively.We also have made a web server for researchers to use, which is available at http://cal.tongji.edu.cn/TurnP/index.jsp.Conclusions: All of the results were outstanding for turn predictions and confirmed the good performance of TurnP for practical applications.We introduced two novel features: SPSSM and shape string profiles, which can accurately reflect the characteristics of turns.TurnP is a pioneering work for the prediction of turns throughout a protein ; we believe it can be complementary to other protein secondary structure prediction methods and may be useful for protein three-dimensional structure prediction .