Fig. 1

An Overview of GenSET Phase 1 selection of the training and testing sets for T3SS effector prediction. Protein or nucleotide sequences from each genome were grouped into three categories that included (i) all known T3SS effectors, (ii) non-effectors including non-T3SS annotated proteins, and (iii) all unannotated hypothetical proteins including all T3SS-related proteins. Fifteen randomly picked effectors (E. coli, S. dysenteriae, and S. Typhimurium) or 21 effectors (P. syringae) from (i) became the positive set. The negative training set was 10-fold larger group of non-effector randomly selected from (ii) of the same genome. GenSET was trained on the positive and negative sets (training set) using unfiltered attributes and filtered attributes and then applied to all remaining sequences of the whole genome (testing set)