An improved prediction of catalytic residues in enzyme structures

Abstract

The protein databases contain a huge number of function unknown proteins, including many proteins with newly determined 3D structures resulted from the Structural Genomics Projects. To accelerate experiment-based assignment of function, de novo prediction of protein functional sites, like active sites in enzymes, becomes increasingly important. Here, we attempted to improve the prediction of catalytic residues in enzyme structures by seeking and refining different encodings (i.e. residue properties) as well as employing new machine learning algorithms. In particular, considering that catalytic residues can often reveal specific network centrality when representing enzyme structure as a residue contact network, the corresponding measurement (i.e. closeness centrality) was used as one of the most important encodings in our new predictor. Meanwhile, a genetic algorithm integrated neural network (GANN) was also employed. Thanks to the above strategies, our GANN predictor demonstrated a high accuracy of 91.2% in the prediction of catalytic residues based on balanced datasets (i.e. the 1:1 ratio of catalytic to non-catalytic residues). When the GANN method was optimally applied to real enzyme structures, 73.9% of the tested structures had the active site correctly located. Compared with two existing methods, the proposed GANN method also demonstrated a better performance.

Wavelet transform analysis of NMR structure ensembles to reveal internal fluctuations of enzymes

“…Highly sequenced conservation is one of the most significant features of catalytic residues (Tang et al 2008). Conversation scores ranging from 1 (most variable) to 9 (most conserved) were assigned using the Consurf Server (Landau et al 2005).…”

“…Then we picked the most conserved ones from the first selection due to the highly conservation of catalytic residues (Bartlett et al 2002;Tang et al 2008). Finally, we ranked the selected sites by their mean solvent accessibility surface areas.…”

“…This result was also established in other applications (Chen and Bahar 2004;Yang and Bahar 2005;Yang et al 2009). Tang et al reported that closeness centrality plays an important role in predicting catalytic residues, and active sites usually have more interactions with other residues (Tang et al 2008). The interactions between catalytic residues and the other residues restrain their fluctuations.…”

Multivariate analysis of properties of amino acid residues in proteins from a viewpoint of functional site prediction

“…Other methods HOMO-LUMO ± 10 [13] 39.3 92.9 7.0 THEMATICS Z = 0.99 [9] 50.4 17.9 THEMATICS-SVM [10] 61.0 96.9 20.0 Gutteridge et al [17] 68.0 16.0 SARIG [18] 57.4 91. 9 8.1 GANN [20] 73.2 96.2…”

“…Ota et al successfully developed a method for combining sequence and structural similarity data with the protein stability change accompanied by mutation [15]. Although similar attempts have already been done by several groups [16][17][18][19][20], there is room for further investigation.…”