The CoMentG server allows to interactively browse the relationships between biomedical entities inferred from co-mentions in the scientific literature. The procedure for inferring these relationships is described in detail in:
Please cite that reference when reporting any data or result obtained from this server.
The access to this server is free to everyone.
This web site contains the results of the application of the CoMent methodology to the detection of different types of relationships between terms representing diseases (DOID and MONDO ontologies), biological processes, molecular functions and cell compartments (GO), clinical signs (HPO), cell types (CL ontology), tissues and body parts (UBERON), bacteria (MeSH terms related to microorganisms) and chemical compounds (human-related metabolites from HMDB and MeSH terms related to chemical compounds).
From the DOID disease ontology we retrieved 10,949 terms representing different diseases. From the whole list of MONDO terms, we exclude those linked to HPO terms (see below) in an attempt to avoid those representing symptoms, ending up in a final list of 21,926 terms. From Gene Ontology (GO) we retrieved the terms from the three subontologies ("molecular function", "biological process" and "cellular compartment" associated to at least one human gene in the GOA annotations, in an attempt to restrict to human-specific GO terms, obtaining a final list of 18,892 terms. For the Human Phenotype Ontology (HPO), we took all the terms that have “phenotypic abnormality” (HP:0000118) as an ancestor in the HPO hierarchy, in order to avoid terms not related to phenotypes or clinical signs, ending up in a final list of 16,218 terms. For the CL ontology, we took all CL:XXXXX terms, that are those representing different cell types (2,532 terms). For the UBERON ontology we took all UBERON:XXXXX terms, representing different tissues and body parts (14,273 terms). From the C and D subsets of the MeSH vocabulary, we retrieved the terms under the "B03" category of the hierarchy as well as those associated to the "semantic type" T007, in an attempt to get the terms of that vocabulary representing microorganisms. We impose the additional constraint that the terms must be linked to a NCBI TaxID, ending up in a final list of 19,194 terms. From MeSH we also retrieved the terms representing chemical compounds as those under different "semantic types" indicative of that, such as T109, T116, T121, etc (2,915 terms). Finally, from the Human Metabolite Database (HMDB) we retrieve all entries (205,011). Nevertheless, contrary to the other datasets, most of these chemical compounds are never mentioned in PubMed, and hence our final list (HMDB compounds mentioned in PubMed) contains 24,880 terms. For all these datasets, we retrieved the terms' names and synonyms annotated in the corresponding fields of the resources. For HMDB we excluded any synonym associated to more than one compound, as some generic names are annotated as "synonyms" in that resource.
With all these terms (and their synonyms) we perform the searches in PubMed as described in the reference above. All the data retrieval and the bibliographic searches were performed between April and October 2022. With the lists of PubMed entries mentioning each term, we detect the significant co-mentions with the procedure described in the reference above. This procedure led to a final list of 7,060,992 co-mention relationships (p-val<=0.001) between terms of those datasets.
For general instructions on how to use this web interface as well as a guided tutorial go to: