Online Supporing Infomation A:
The accession numbers and mature piptide of the 261 toxin sequences and 60 negative dataset in the benchmark dataset classified into 5 types (subsets): (1) A-conotoxin superfamily 63, (2) M-conotoxin superfamily 48, (3) O-conotoxin superfamily 95, (4) T-conotoxin superfamily 55, and (6) Negative dataset 60. None of proteins included here has more than 20% pairwise sequence identity to any other in a same subset. See the text of the paper for further explanation. To download the data in the Online Supporting Information A, click Supp-A.