An in silico screening method for novel dipeptides from a dipeptide training set was performed. A 3D pharmacophore
model with three H-bond acceptor projections (Acc2), and one H-bond donor projection (Don2) was obtained. To
validate the pharmacophore model, a test set containing 10 reported dipeptides was screened and all were found to be bioactive.
Eleven novel dipeptides (IF, GD, DA, TE, TA, ES, SS, ST, SD, QD, QE) from the silkworm pupae protein peptide
database were predicted to have ACE inhibitory bioactivity. The interaction mechanisms of the dipeptides with the ACE
active pocket were elucidated by molecule docking, and besides involving interaction bonds, the interaction force equipoise
of the peptides was also an important factor in their bioactivity.