Skip to main content

Overview of Literature

Updated

Source:

  • Except for citation data, all data are sourced directly from PubMed, PubMed Central (PMC) XML downloads, and preprint servers*. 
  • Citation data are sourced from the NIH Open Citation Collection (OCC). The OCC data underlie the NIH iCite analytic platform and include citation data merged from public sources as well as a machine learning pipeline that extracts, resolves, and disambiguates references from full-text articles. To read more about the NIH Open Citation Collection see this PLOS Biology paper. 

*Most literature in PubMed is peer-reviewed; however, the level of peer-review varies for different journals, and not all journals in PubMed are peer-reviewed journals. In addition, PubMed is piloting inclusion of preprints, which will appear under the pub type of preprint. Read more about the peer-reviewed journals included in PubMed and the preprint pilot. 

 

Preprint sources:

Preprint articles are sourced from the following preprint servers:

arXiv, bioRxiv, ChemRxiv, Focusarchive, medRxiv, MetaArXiv, PsyArXiv, Preprints.org, Qeios, Research Square and SocArXiv

 

Update Frequency:

 

Source Updates
PubMed Daily except for citation data, which are updated monthly
Preprints Preprint data are updated weekly

 

Enrichments:

  • Admin IC, Activity Code, and Core Project Number searchability
  • Citation data (RCR, citations, cited by)
  • Translational science (cited by clinical article, is clinical article, animal percent, human percent, molecular/cellular percent, APT score)
  • Organizations are normalized to their preferred form
  • Enhanced author name search that better handles author variations
  • iCite article type
  • Condition, Target, Chemicals & Drugs, Devices, MeSH Plus are extracted from Title, Abstract, MeSH and keyword fields using synonyms
  • See the Literature glossary for more detail

Linking:

Refine Filters Panel:  Has Linked Data  Linkage Description
Grants

Linking for Literature supported by HHS Grants comes from SPIRES, PubMed data, and PI Disambiguation algorithms.

  • SPIRES match case 3.5 and higher are used. Match case 3.5 is a match case 3 with a name match between the grant PI and the authors on the paper.
  • PubMed grants links are normalized using NLP techniques and grant number extraction algorithms. If necessary, remaining ambiguity from grant numbers are removed via matching PI names.
  • Disambiguation links are created via OPA's customized PI disambiguation algorithm and filtered by year.
Patents Linking for Literature cited by patents comes from the Non-Patent Citation field in USPTO data. A citation resolution algorithm resolves a free-text citation to a PMID for the linking.
Clinical Trials Linking for Literature cited by, or having the results of, a clinicaltrials.gov clinical trial.

Learn more about each of these links in the View Linked Data article.