Fig. 11

The p-value of the largest fragment between every trypsin-elastase and trypsin-chymotrypsin pair in Dataset C was estimated using training data generated at .0625 Å (blue line), .125 Å (red line), and .250 Å (yellow line). The vertical axis plots p-value, fragments are sorted by ascending p-value along the horizontal axis. The black line indicates the \(\alpha\) threshold of 0.02. Fragments with p-values lower than 0.02 result in a rejection of the null hypothesis and thus a prediction that the elastase or chymotrypsin in the pair has binding preferences that are different from that of trypsin. p-values above the threshold do not result in a rejection of the null hypothesis and thus the incorrect presumption that an elastase or chymotrypsin has binding preferences similar to trypsin. The finer resolution training data made possible with pClay yielded more accurate predictions