RandomPairs data set
Previous studies have proven that pairwise proteins do not interact if one protein from these two proteins reverses the order of amino acids. However, selecting the appropriate length of the artificial protein sequence is difficult. Thus, the PPIs from the positive data set are shuffled to obtain non-interacting pairs and ensure that the length of the proteins remains the same to address the shortcomings of the first strategy. The data set was called RandomPairs data set.
RecombinePairs data set
Two PPIs randomly selected from the positive data set are labeled PA1PB1 and PA2PB2. Then, proteins PC1 and PC2 are selected from {PA1, PB1} and {PA2, PB2} separately to combine a new non-interacting PC1PC2 pair. The most important requirement for the second strategy is that the generated non-interacting pairs must not appear in the positive data set. The final data set is called the RecombinePairs data set.
All Rights Reserved Copyright @ 2015|Jiancang Zeng
Last Modified in 2015/2/8