Proteome-Wide Analysis of Human Deletions.

Protein deletions are frequent among both natural and pathogenic variations. Many of them are misclassified in variation databases and the literature. Nonsense-mediated decay prevents the expression of many nucleotide deletions. Many variants classified as protein deletions are not expressed at all. We conducted an exhaustive systematic analysis of three types of deletions: N- and C-terminal deletions, as well as internal deletions within protein sequences. In addition, we compared natural and pathogenic internal deletions. We collected an extensive dataset of reliable deletions in many proteins and then performed extensive statistical analyses to investigate properties of deletions and proteins that contain them. We studied the properties of protein deletions, including deletion length and position, amino acid composition, flanking amino acid sequence context, the functions and properties of deletion-containing proteins, the functional roles of the deleted regions, the positioning within protein domains and protein structure, as well as sequence conservation and involvement in protein-protein interaction networks. We found several statistically significant differences between the deletion types and between benign and pathogenic deletions. The obtained insight can be used, for example, for variation interpretation, prediction method development, and analysis of variation mechanisms and effects.