Analysis of fumarate-sensitive proteins and sites by exploiting residue interaction networks.

Fumarate adduction to cysteines has been implicated in the pathogenesis of several disorders. Its role, however, still remains elusive, and the need of predictive methods has not yet been met. The reactivity of cysteines found in fumarate-sensitive proteins was predicted when the collected data for eight network-type features were analyzed using classification models. Therefore, methods for evaluating the likelihood of a cysteine site to be modified by fumarate could be developed by combining concepts of network theory and machine learning.