EVEX: a PubMed-scale resource for homology-based generalization of text mining predictions

In comparative genomics, functional annotations are transferred from one organism to another relying on sequence similarity. With more than 20 million citations in PubMed, text mining provides the ideal tool for generating additional large-scale homology-based predictions. To this end, we have refined a recent dataset of biomolecular events extracted from text, and integrated these predictions with records from public gene databases. Accounting for lexical variation of gene symbols, we have implemented a disambiguation algorithm that uniquely links the arguments of 11.2 million biomolecular events to well-defined gene families, providing interesting opportunities for query expansion and hypothesis generation. The resulting MySQL database, including all 19.2 million original events as well as their homology-based variants, is publicly available at http://bionlp.utu.fi/.In comparative genomics, functional annotations are transferred from one organism to another relying on sequence similarity. With more than 20 million citations in PubMed, text mining provides the ideal tool for generating additional large-scale homology-based predictions. To this end, we have refined a recent dataset of biomolecular events extracted from text, and integrated these predictions with records from public gene databases. Accounting for lexical variation of gene symbols, we have implemented a disambiguation algorithm that uniquely links the arguments of 11.2 million biomolecular events to well-defined gene families, providing interesting opportunities for query expansion and hypothesis generation. The resulting MySQL database, including all 19.2 million original events as well as their homology-based variants, is publicly available at http://bionlp.utu.fi/.C