#2 Disambiguate gene mentions using already normalized genes from the same text

[Jörg Hakenberg]

When a gene ends up with only candidate IDs but not final decision, try to resolve the ambiguity by looking at other genes in the same text. If some other gene mentions have successfully been narrowed down to just one ID, use this information for the ambiguous mention.

• if only one of the candidate IDs was found as the final ID for one or more other gene mentions, use it to normalize the ambigious gene
• could use position information: same sentence, or within 3 characters (indicates parenthesis)

Example:
pmid start stop mention tool candidateIds finalId taxId score
20818844 33 36 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 127 163 serine-threonine protein kinase B-RAF GNATGM 207;2185;673;9874 null N/A N/A N/A
20818844 166 169 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 410 413 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 808 811 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 978 981 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 1278 1281 BRAF GNAT 6635 673 BRAF 9606 3.0
20818844 1696 1699 BRAF GNAT 6635 673 BRAF 9606 3.0