Skip to main content
. 2018 Jan 11;8:546. doi: 10.1038/s41598-017-18699-8

Figure 2.

Figure 2

Developing a bioinformatic and systematic literature review for defining proteomes. The entire body of Web of Science literature was searched. A series of terms were used to select a smaller population of publications of interest and the abstract text for each paper was then downloaded (ad). A Perl algorithm was developed to search the downloaded abstract text for: at least one protein name as stated in Uniprot, “skin,” “Immunohistochem*”, and/or “Mass Spec*” (e) and created an Excel file with the list of DOIs with the contained terms (f). The papers in the Excel list were then manually reviewed for evidence (IHC, MS, ELISA, WB or zymography) of the protein in healthy human skin (g). Proteins were added to the database with captured data on: protein name, accession number, and gene name (from Uniprot) and; from target articles: skin sample information (age, sex, and anatomical site), localisation within the tissue and any associated pathologies.