Coverage of the Antibody Registry

For Figure ​Figure3,3, the list of human proteins was obtained from UniProt (RRID:SCR_002380), where the protein (a) had evidence of its existence at the protein level, (b) was reviewed and (c) was a human protein. Following these requirements, the following query was constructed: https://www.uniprot.org/uniprotkb?query=(existence:1)%20AND%20(reviewed:true)%20AND%20(organism_id:9606), which was run on 12 May 2020. For each protein name (only taking the first name if there are synonyms), the Antibody Registry index (available at https://scicrunch.org/resources/data/source/nif-0000-07730-1/search) was queried for each protein name on 12 August 2022, and the number of results returned for each query was recorded (see supplemental data file). The top 30 genes with the largest number of results were manually inspected to remove cases where the result count did not accurately reflect the true number of antibodies, e.g. the gene ‘CAT’ matching to all entries with ‘Cat#’. For cases where gene names could potentially overlap with other terms, as determined by a human curator, the number of results were replaced with the results of a manual query to the Antibody Registry index where the ‘Target Antigen’ field was required to match the gene name. While this underestimated the true number of antibodies as gene names or synonyms are also found in the antibody name field, it was preferred for such cases where the default query is a significant overestimate.

Percentage of the proteome that is covered by antibodies. This cumulative plot shows that about 10% of the total proteins have less than 15 antibodies, about 10% of proteins have more than 100 antibodies, but roughly 80% of known human proteins have 15 to about 100 antibody reagents available for researchers. While we do not know if there are high quality antibodies validated for each target, we do at least know that most of the proteome is accessible by these key reagents. For a listing of protein names and the associated counts please see the supplementary data file.

Information about the use of RRIDs in the scientific literature is gathered by curators weekly using the SciBot semi-automated workflow described previously (11; Section ‘Corpus of papers containing RRIDs’). The data is made freely accessible via the Hypothes.is API (RRID:SCR_000430). To obtain statistics on usage of Antibody Registry RRIDs in the literature, we queried a local version of the Hypothes.is database (last updated on Aug 8, 2022) to return counts of all RRIDs that include ‘RRID:AB_’, but are not marked as duplicates. These data can be viewed in Hypothes.is at https://hypothes.is/users/SciBot, as annotations on the individual papers (https://hyp.is/aVKDGt_PEeu4jjuX5OipXQ/www.ncbi.nlm.nih.gov/pmc/articles/PMC8250850/), on resolver pages as papers associated with a particular antibody record (e.g. https://n2t.net/RRID:AB_2314866), or in a computationally accessible way by adding ‘.json’ as suffix to the RRID (e.g. https://n2t.net/RRID:AB_2314866.json).