Skip to main content
. 2021 Apr 27;17(5):3221–3232. doi: 10.1021/acs.jctc.1c00159

Figure 12.

Figure 12

Self-assembling peptides with high AP score with log P < −4 found via active learning; in each case, the models find peptides above the maximum and mean values from the random set (red and blue lines, respectively). The trend seems to be at a restriction of this magnitude; the larger the data set, the better the model is able to learn.