SKEMPI 2.0: An updated benchmark of changes in protein-protein binding energy, kinetics and thermodynamics upon mutation

Justina Jankauskaite; Brian Jiménez-García; Justas Dapkunas; Juan Fernández-Recio; Iain H. Moal

Another source of variation is the origins of the interacting proteins. While entries range from viral and bacterial to the higher eukaryotes, biases are evident in the over-representation of model organisms including humans. With the exception of the pMHC-TCR, antibody-antigen and protease-inhibitor classes, most of the interactions are endogenous. Nevertheless, the set also includes exogenous interactions ranging from those between proteins from different individuals within the same species, namely the sex fusion proteins Juno and Izumo1 from human sperm and egg respectively [4] , to host-pathogen interactions such as adenovirus and coronavirus interactions with human receptors during viral entry [65] , [27] , to the inhibition of acetylcholinesterase by the snake venom neurotoxin fasciculin [2] . For the pMHC-TCR interactions there are a variety of presented antigens, including exogenous viral peptides and gluten, as well as endogenous autoimmune and cancer peptides. The antibody interactions include pathogen antibodies, as well as antibodies raised and optimised to target extracellular therapeutic targets. The protease interactions are mostly exogenous, arising from their inherent cross-reactivity due to the convergent evolution of their canonical inhibitory loop.