Figure 1
From: Rule-based knowledge aggregation for large-scale protein sequence analysis of influenza A viruses
Retrieval performances of the NCBI nucleotide and protein databases. Each chart shows 5 pairs of bars, one for each extracted property. The first (darker) bar of each pair shows the performance for the GenBank database while the second (lighter) bar shows the value for GenPept. The first chart shows the percentage of source documents from which a property value could be extracted, while the second graph shows the percentage of accurate values extracted, measured against the manually annotated dataset.