Bioinformatics Advance Access published March 22,

Khatun et al. (2004) developed contact potentials and showed a correlation of 0.46 in a dataset of 1356 mutations. Capriotti et al. (2005) developed SVM based method, which predicts protein stability with 80 % accuracy using structural information and 77%

Bioinformatics Advance Access published March 22, 2007

Khatun et al. (2004) developed contact potentials and showed a correlation of 0.46 in a dataset of 1356 mutations. Capriotti et al. (2005) developed SVM based method, which predicts protein stability with 80 % accuracy using structural information and 77%

L.-T. Huang et al.

Khatun et al. (2004) developed contact potentials and showed a correlation of 0.46 in a dataset of 1356 mutations. Capriotti et al. (2005) developed SVM based method, which predicts protein sta-bility with 80% accuracy using structural information and 77% from sequence. Cheng et al. (2006) also used SVMs and reported the accuracy of 84% using structure information. Parthiban et al. (2006) proposed a method based on torsion and atom potentials, which predicted the stability with more than 80% accuracy. The present method could discriminate the stabilizing and destabilizing mutants at an accuracy of 82% and predict the protein stability changes upon mutations with the correlation of 0.70 from amino acid sequence. This analysis shows that the performance of our method is similar to or better than other methods in the literature. The main features of the present method are: (i) it is based on the neighboring residues of short window length, (ii) it can predict the stability from amino acid sequence alone, (iii) developed dif-ferent servers for discrimination and prediction, and integrated them together, (iv) utilized the information about experimental conditions, pH and T, and (v) implemented several rules for dis-crimination and prediction from the knowledge of experimental stability and input conditions: (i) if the deleted residue is Ala and the neighboring residues contain Gln, then the predicted stability change will be negative (accuracy = 97.1%), (ii) if the deleted residue is Glu and its second neighbor at N-terminal is Met, the mutation stabilizes the protein (accuracy = 100%) and (iii) if the deleted-residue belongs to Y, W, V, R, P, M, L, I, G, F or C, and the introduced-residue belongs to T, S, P, K, H, G or A, then the predicted stability change will be -2.05 kcal/mol (mean absolute error = 1.57 kcal/mol). Additional rules are provided on the web.

Fig. 1. Snapshot showing the necessary items to be given as input for discrimination and prediction.

4 SERVER DESCRIPTION

The input options for discrimination/prediction are shown in Fig-Fig 2. The results obtained for predicting the stability change along with

ure 1. The program takes the information about the mutant and the related information of neighboring residues. mutated residues, three neighboring residues on both sides of the

Cheng, J., Randall, A. and Baldi, P. (2006) Prediction of protein stability changes for mutant residue along with pH and T. In the output, we display the

single-site mutations using support vector machines, Proteins, 62, 1125-1132.

predicted protein stability change upon mutation along with input Freund, Y. and Schapire, R.E. (1997) A decision-theoretic generalization of on-line conditions (Figure 2). In the case of discrimination, we show the learning and an application to boosting, Journal of Computer and System Sci-ences, 55, 119-139. effect of the mutation to protein stability, whether stabilizing or

Gilis, D. and Rooman, M. (1996) Stability changes upon mutation of solvent-destabilizing. Both discrimination and prediction services offer an

accessible residues in proteins evaluated by database-derived potentials, J. Mol.

option for additional sequence composition information of

Biol., 257, 1112-1126.

neighboring residues (Figure 2). The bar chart shows the number Gromiha, M.M., An, J., Kono, H., Oobatake, M., Uedaira, H. and Sarai, A. (1999) of amino acids of each type. The two pie charts below represent ProTherm: Thermodynamic Database for Proteins and Mutants, Nucleic Acids

Res., 27, 286-288. the percentage of residues according to polarity and the metabolic

Gromiha, M.M., Oobatake, M., Kono, H., Uedaira, H. and Sarai, A. (1999) Role of role of amino acids.

structural and sequence information in the prediction of protein stability changes:

In addition, we have provided the datasets used in the present comparison between buried and partially buried mutations, Protein Eng., 12, 549-work along with the references and links to related web servers. A 555.

Guerois, R., Nielsen, J.E. and Serrano, L. (2002) Predicting changes in the stability of help page is also provided for the details to be given in the input.

REFERENCES

Bava, K.A., Gromiha, M.M., Uedaira, H., Kitajima, K. and Sarai, A. (2004)

ProTherm, version 4.0: thermodynamic database for proteins and mutants, Nucleic Acids Res., 32, D120-121.

Breiman, L. (1984) Classification and regression trees. Wadsworth International

Group, Belmont, Calif.

Capriotti, E., Fariselli, P. and Casadio, R. (2004) A neural-network-based method for

predicting protein stability changes upon single point mutations, Bioinformatics, 20 Suppl 1, I63-I68.

Capriotti, E., Fariselli, P. and Casadio, R. (2005) I-Mutant2.0: predicting stability

changes upon mutation from the protein sequence or structure, Nucleic Acids Res., 33, W306-310.

proteins and protein complexes: a study of more than 1000 mutations, J. Mol. Biol., 320, 369-387.

Huang, L.-T., Saraboji, K., Ho, S.-Y., Hwang, S.-F., Ponnuswamy, M.N. and

Gromiha, M.M. (2007) Prediction of protein mutant stability using classification and regression tool, Biophysical Chemistry, 125, 462-470.

Khatun, J., Khare, S.D. and Dokholyan, …… 此处隐藏：4292字，全部文档内容请下载后查看。喜欢就下载吧 ……