Benchmark Data
Online Supporting Information A. The learning dataset contains 653 Gram-negative bacterial proteins classified into 8 subcellular locations according to the experimental annotations. Both the accession numbers and sequences are given. None of the proteins has more than 25% sequence identity to any other in the same subset (subcellular location). See the reference given on the top page of the web-server for further explanation. Click Supp-A to download the dataset.

Online Supporting Information B. The testing dataset contains 643 Gram-negative bacterial proteins classified into 8 subcellular locations according to the experimental annotations. Both the accession numbers and sequences are given. None of the proteins has more than 25% sequence identity to any others in the same subcellular location of either the dataset here or the training dataset in the Online Supplementary Materials A. See the reference given on the top page of the web-server for further explanation. Click Supp-B to download the dataset.

Online Supporting Information C. The degenerate testing dataset used for comparing the performance between PSORT-B and the predictor of this paper. The dataset contains 1,114 Gram-negative bacterial proteins classified into 5 subcellular locations: (1) cytoplasm, (2) extracell, (3) inner membrane, (4) outer membrane, and (5) periplasm. Click Supp-C to download the dataset.