To improve the scalability of our literature curation efforts, we have streamlined our literature selection processes and implemented software infrastructure to support automation for these processes. We now store the full text of papers extracted from PDFs downloaded from publishers and assess the relevance of papers using keyword searches. Downloading papers from PLoS journals is performed automatically and takes advantage of the PLoS API’s full text search capabilities. Full text searches improve the identification of relevant papers for MGD because important keywords such as mouse and murine are often not mentioned in article titles and abstracts. In the eleven months following the implementation of the improved literature selection processes (aka, literature triage), individual curator efficiency in identifying papers relevant to mouse phenotypic alleles has increased by 83% as measured by number of relevant papers identified by an individual per unit time. For our user communities, the increased efficiency in literature curation means a shorter time between publication and accessibility of the phenotype and disease annotations from MGD.
