To demonstrate the functionality and output of ANNOVAR, we analyzed the input file shown in Table 1. We applied gene-based annotation procedure using RefSeq gene definitions (15), though the UCSC Genes definition (16) or Ensembl Gene definition (13) can be used alternatively. Two output files were generated, one of which annotates the location of each variant with respect to genes (one variant per output line), that is, whether it is exonic, intronic, intergenic, splicing site, 5′/3′-UTR, upstream/downstream of genes, or whether it has invalid input format. The other output file contains amino acid changes that may be caused by the mutation. We utilized a standardized nomenclature (17) to annotate non-synonymous SNVs and indels on cDNA or on proteins. For example, the first mutation has a functional consequence as NOD2:NM_022162:exon4:p.R702W, indicating that the mutation causes a non-synonymous change in exon 4 of the NOD2 gene. Since each gene may have multiple splicing isoforms in RefSeq annotations, the RefSeq transcript identifiers are always given after the gene name, and some variants may be annotated with respect to multiple alternative transcripts.
