Automated transfer of experimentally-verified GO annotation data to close orthologs

Sascha Steinbiss, GeneDB curators; 2015

This reference is used to describe functional annotations transferred from one or more reference (“source”) organisms to a newly annotated (“target”) organism on the basis of ortholog cluster membership. In detail, predicted (e.g. by AUGUSTUS, see doi:10.1186/1471-2105-7-62) or transferred (e.g. via RATT, see doi:10:1093/nar/gkg1268) gene models in the target genome are translated and processed by OrthoMCL 1.4 together with reference protein sequences to produce clusters of gene products derived from orthologous genes. For each cluster, GO terms are automatically transferred from source products to the target gene products if they are experimentally verified (IDA (ECO:0000314), IMP (ECO:0000315), IPI (ECO:0000353), IGI (ECO:0000316), (EXP ECO:0000269). They are tagged with the ISO evidence code and the “with/from” is populated with the source feature references (e.g. “GeneDB:LmjF.28.0960”). OrthoMCL runs are done using the parameterization suggested in the OrthoMCL algorithm document (blastall -F ‘m S’ -e 1e-5).