Beyond MitoCarta-expanding the list of candidate proteins involved in mitochondrial functions using a biological...

Mitochondrial diseases are the result of pathogenic variants in genes involved in the diverse functions of the mitochondrion. A comprehensive list of mitochondrial genes is needed to improve gene prioritization in the diagnosis of mitochondrial diseases and development of therapeutics that modulate mitochondrial function. MitoCarta is an experimentally derived catalog of proteins localized to mitochondria. We sought to expand this list of mitochondrial proteins to identify proteins that may not be localized to the mitochondria yet perform important mitochondrial functions. We used a computational approach to assign statistical significance to the overlap between STRING database gene network neighborhoods and MitoCarta proteins. Using a data-driven stringent significance threshold, 2059 proteins that were not located in MitoCarta were identified, which we termed mitochondrial proximal (MitoProximal) proteins. We identified all of the oxidative phosphorylation complex subunits and 90% of 149 genes that contain confirmed oxidative phosphorylation disease causal variants, lending validation to our methodology. Among the MitoProximal proteins, 134 are annotated to be localized to mitochondria but are not in the MitoCarta 3.0 database. We extend MitoCarta nearly 3-fold, generating a more comprehensive list of mitochondrial genes, a resource to facilitate the identification of pathogenic variants in mitochondrial and metabolic diseases.