PRRs sequences (positive data) were obtained from the database PRRDB2.0. Initially the total PRRs taken were 2727, which were reduced to 179 unique PRRs after removal of identical sequences.
Negative dataset
The negative dataset was created by collecting random sequences from Swiss-Prot which were not PRRs. This data was further filtered to obtain sequences which were non-redundant amongst themselves. The negative dataset finally constituted of 274 non-PRR sequences.