GtoPdb NAR database issue 2018: Journal to database connectivity and journal to GtoPdb links

The following blog post acts as supplementary data to the 2018 NAR Database Issue

Journal to Database connectivity

The citation provenance of all entity records and contextual comments selected by the curators and NC-IUPHAR members in GtoPdb is supported by four document types. These are journal papers with PubMed Identifiers (PMIDs, 30,894), journal papers without PMIDs (246), book references (72) and patent numbers (412). We also have 109 URL-only citations we have judged of good reputation and expected stability (some of which will get displaced when appropriate journal papers appear). The key axis of connectivity that we facilitate is PubChem-to-PubMed reciprocal linking. The importance of this overall has been described by the PubChem team in some detail, including the contribution of GtoPdb as one of the mapping sources [1].

The set of curated ligand references (for quantitative activity data at targets as well as selected ancillary references, such as completed clinical trial reports) form part of the SID records we submit to PubChem, which has a number of linking consequences. Note also that we uniquely specify the explicit location of the ligand structure within the reference. For example, ligand id: 8135 is named “compound 21 [PMID: 23312943]” and can thus be discriminated from no less than seven other “compound 21”s in the database by their specific PMID suffixes. Figure 1 illustrates the link between GtoPdb compounds and “Depositor Provided PubMed Citations” (DPPMC) both in the SID from us and merged in the CID from other submitters. Crucially, this relationship is reciprocal as we can see in the lower panel of Figure 1. This means that any user coming in to the NCBI Entrez system [2], either via PubMed or PubChem, can connect the paper to the structure or vice versa. In this example, we are the only source that has submitted a connection and the structure can be located in the paper (i.e. as compound 21). Conversely, popular compounds (e.g. approved drugs) may have PubMed connections in their CIDs from many submitters, but ours will include the quantitative binding data reference which may be before the drug was awarded an International Nonpropietary Name (INN).

Figure 1.  GtoPdb to PubChem to PubMed connectivity for ligand 8135.

Our overall PubMed statistics are shown in Table 1.

Table 1. GtoPdb PubMed statistics

The majority of PMIDs (22,060) are associated with individual targets as well as commentaries on families, accumulated from curation and committee updates over 14 years. Internally we can attribute 9,673 PMIDs to ligand-specific references. From our 8978 SIDs, 82% have at least one DPPMC making a total of 9,086 PMID links. Of these, 6,011 refer to the quantitative interaction. We have analysed the journal breakdown for our ligands as shown in Figure 2 which reflects our empirical primary, secondary and tertiary reference classifications. For example, primary citations as first reports of binding data between ligands and targets are often selected from the Journal of Medicinal Chemistry, while we generally cite the British Journal of Pharmacology (BJP) in relation to in vivo rodent pharmacology, and occasionally the British Journal of Clinical Pharmacology (BJCP) for clinical trial reports. To discern if there was an immunopharmacological curation signal in our literature we compared Figure 2 with the PMIDs only from GtoImmuPdb. It was interesting to note that for the primary references we selected for quantitative ligand interactions, the overall pattern was similar. Notably, however, Journal of Immunology had moved up from a ranking of 17th in Figure 2 to 6th in the GtoImmuPdb references.

Figure 12.  Top-twenty journals from the 8,756 PMIDs cited in the interaction comments.

Journal-to-GtoPdb links

Our engagement with the BJP in the provision of live out-links has been described previously [3]. The major enhancement for this year is that Wiley have transitioned to in-line links in the text (at first mention), rather than the previous method of adding separate tables to the manuscripts. Taking the recent BJP papers from Volume 174, Issue 18 September 2017 as an example, the 12 papers therein have 134 out-links to GtoPdb. This year has also produced our first “circular” example where GtoPdb team members are co-authors on a Systems Pharmacology study, partly derived from the database for which we have added a set of links “back in” [4]. This year Wiley have also introduced the same GtoPdb out-links for the BJCP.

1. Kim, S., Thiessen, P.A., Cheng, T., Yu, B., Shoemaker, B.A., Wang, J., Bolton, E.E., Wang, Y. and Bryant, S.H. (2016) Literature information in PubChem: associations between PubChem records and scientific articles. J Cheminform, 8, 32. PMID: 27293485

2. Gibney, G. and Baxevanis, A.D. (2011) Searching NCBI databases using Entrez. Current protocols in bioinformatics, Chapter 1, Unit 1 3. PMID: 21975942

3. McGrath, J.C., Pawson, A.J., Sharman, J.L. and Alexander, S.P. (2015) BJP is linking its articles to the IUPHAR/BPS Guide to PHARMACOLOGY. Br J Pharmacol, 172, 2929-2932. PMID: 25965085

4. Benson, H., Watterson, S., Sharman, J., Mpamhanga, C., Parton, A., Southan, C., Harmar, A. and Ghazal, P. (2017) Is systems pharmacology ready to impact upon therapy development? A study on the cholesterol biosynthesis pathway. Br J Pharmacol. [Epub ahead of print] PMID: 28910500