The APPRIS annotation pipeline deploys a range of computational methods to provide value to the annotations of the human genome. The server flags variants that code for proteins with altered structure, function or localization, and exons that are evolving in a non-neutral fashion. APPRIS also selects one of the CDS for each gene as the principal functional isoform. The pipeline is made up of separate modules that combine protein structure and function information and evolutionary evidence. Each module has been implemented as a separate web service.firestar (Lopez et al. 2007, 2011) is a method that predicts functionally important residues in protein sequences.Matador3D is locally installed and checks for structural homologs for each transcript in the PDB (Berman et al. 2000).SPADE uses a locally installed version of the program Pfamscan (Finn et al. 2010) to identify the conservation of protein functional domains.INERTIA detects exons with non-neutral evolutionary rates. Transcripts are aligned against related species using three different alignment methods, Kalign (Lassmann and Sonnhammer 2005), multiz (Blanchette et al. 2004), and PRANK (Loytynoja and Goldman 2005), and evolutionary rates of exons for
