Additional file 1 of Protein secondary structure prediction using a small training set (compact model) combined with a Complex-valued neural network approach

Table S1. The 25 sequences of the G Switch Proteins dataset (GSW25). The 12 G A sequences and 13 G B sequences are given and cited with their original source. Table S2. The 55 proteins of the compact model (SSP 55). The protein names, SCOP classes, folds, number of residues, and the Q 3 achieved per protein are given. Table S3. The confusion matrices broken down by SCOP classes, are given for the SSP 55 proteins. Table S4. The confusion matrices broken down by SCOP classes, are given for the remainder of the CB513 dataset (330 proteins). (XLSX 30 KB)