Tips:We recommend that users upload less than 100 data at a time. If there is no feedback or the page is lost, please try to reduce the amount of samples appropriately.
We have presented EP3, an ensemble predictor to distinguish amino acids sample Type III Secreted Effectors from non-T3SEs. Our webserver contains two models, which were trained from training dataset 1 and training dataset 2 and were called EP3_1 and EP3_2, respectively. Based on the species of test dataset, the users can choose the model they want. Tips: The samples of training dataset 1 were mainly derived from 16 species (including P. syringae, Escherichia coli, Salmonella enteric and other species.), while the positive samples of training dataset 2 was cross-reference of 62 species (The proportion of Pseudomonas syringae and chlamydla trachomatis was 17.41% and 10.55%, respectively.).
1. Composition of parallel correlation pseudo-amino acid composition (PC-PseAAC) were used to extract features. 2. Label Propagation Algorithm (LPA) identified the T3SEs from non-T3SE according to the PC-PseAAC.
1. Composition of Distance Pair were used to extract features. 2. Label Propagation Algorithm (LPA) identified the T3SEs from non-T3SE according to the Distance Pair.
1. Composition of distance-based top-n-gram (DT) were used to extract features. 2. Label Propagation Algorithm (LPA) identified the T3SEs from non-T3SE according to the DT.
The improved Smith Waterman Algorithm was utilized to capture the similarity matrix (BLOSUM 35), and Support Vector Machine (SVM) distinguished T3SEs from non-T3SEs
The improved Smith Waterman Algorithm was utilized to capture the similarity matrix (BLOSUM 40), and Support Vector Machine (SVM) distinguished T3SEs from non-T3SEs
The improved Smith Waterman Algorithm was utilized to capture the similarity matrix (BLOSUM 45), and Support Vector Machine (SVM) distinguished T3SEs from non-T3SEs
Supported data file formats are Fasta.
If you have any problems,please contact me.
lijingtju@foxmail.com