Figure 1.

Sample Abstract and "Noisy" Gene List. Underlined and bold words in the PubMed Abstract are the genes were found in the text. The answer key consists of four columns. Column 1 is the file name; column 2 has the model organism unique database identifiers. Column 3 shows whether the gene was found automatically in the abstract (Y), not found and pruned (N), or added by hand (X). Column 4 shows the final set of genes in the answer key. This answer key shows that two genes were given by the database curators (FBgn0000592, and FBgn0002722); the first one was found in the abstract, the second one was not. The third gene (FBgn0026412) was found by our annotators based on the guidelines.

Colosimo et al. BMC Bioinformatics 2005 6(Suppl 1):S12   doi:10.1186/1471-2105-6-S1-S12