Investigation of essential proteins is significantly valuable for understanding of cellular life, drug design and other practical purposes. In most of current studies, essential proteins are generally mined in protein-protein interaction (PPI) networks with diverse topology features. In this study, we investigate what kind of proteins is inclined to be essential from a new perspective. The investigation implies that protein essentiality is correlated with protein domains, which are functional, structural and evolutionary units of proteins. Proteins with a larger Number of Domain Types (NDT) tend to be essential. The analyses on 22 species show that essential proteins identified by NDT are much more than those identified by ten random identifications. The consideration of the structural feature makes us less dependent on network data and thus enables us to investigate protein essentiality of more species with incomplete and/or inconsistent network data.
Keywords: PPI; bioinformatics; domain types; essential proteins; incomplete data; inconsistent data; protein–protein interaction networks.