Fig. 3From: Text-mining clinically relevant cancer biomarkers for curation into the CIViC databasea A Shiny-based web interface allows for easy exploration of the CIViCmine biomarkers with filters and overview pie charts. The main table shows the list of biomarkers and links to a subsequent table showing the list of supporting sentences. b The entirety of PubMed and PubMed Central Open Access subset were processed to extract 87,412 biomarkers distributed between the four different evidence types shown. c Protein-coding variants extracted for each evidence item are compared against somatic variants in COSMIC and > 1% prevalence SNPs in dbSNPBack to article page