DAMpred for Recognizing Disease-Associated SNP Mutations

DAMpred is a method for predicting Disease-associated mutations of proteins in the human genome. A novel training method integrates Bayes classification and artificial neutral network to comprehensively consider sequence-based, biological assembly-based and I-TASSER model-based features and their corresponding probability density on disease-associated dataset and neutral dataset, respectively.

Figure 1 shows the Flowchart of DAMpred for disease-associated mutation prediction. All the features of DAMpred are derived from protein sequence, so its ability to deliver good predictions without experimental structure high-quality extend the application of disease-associated mutation prediction.

Figure 1: The flowchart of DAMpred method.

The DAMpred algorithm was tested on a set of 10,634 mutations from 2,154 proteins. After excluding homologous templates with sequence identity >30% to the target, I-TASSER was able to build correct structure model with a TM-score > 0.5 for 1,863 proteins. Using the I-TASSER model, DAMpred classifies all mutation into two groups, neutral and disease-associated, with an ACC 0.800 and MCC 0.601 in the protein-level 10-fold cross-validation. DAMpred improve 17% on average MCC, indicating that DAMpred give more balanced accuracy, especially for neutral mutations, compared with others methods.