Skip to main content

Table 3 Performances of the two-class PaleAle and PaleAle_H compared with a number of recent methods on the Manesh dataset [39].

From: Accurate prediction of protein secondary structure and solvent accessibility by consensus combiners of sequence and structure information

Method  
PaleAle_H 86.0%
PaleAle 79.2%
NETASA [15] 70.3%
[21] 78.1%
PP [47] 78.1%
PredAcc [11] 70.7%
JNET [13] 75.0%
ACCpro [16] 77.2%
SABLE [17] 77.6%
  1. Performances of the various methods from [21]. The class threshold is 25% for all methods. Templates up to 95% adopted by PaleAle_H.