I assume thousands of scientists reading papers struggle with deciphering coded information, esp. at first encounter. So inefficient/frustrating...
Well the problem with standards is in general that there are too many of them. But we do of course have resources for biological terms that try to combine the various controlled vocabulary and ontology initiatives. The two main ones are NCBO (the US center for Biomedical ontologies) at http://www.bioontology.org/ and the ontology service page at EBI.
In practice we have many (partially) overlapping ontologies, and that is often for a good reason. The same things just need slightly different details or descriptions depending on the field you are using them for. Which means we cannot end up with one big ontology and are more in need of mappings between ontologies. Nevertheless such meta ontologies are also developed, and contain a lot of mapping internally. The experimental factor ontology (EFO) is a good example. These meta ontologies can never be perfect because we also want people to be able to create new ontologies for new purposes and add terms to existing ontologies. We actually need better tools to make such new terms, that are often added to the meta ontologies, propagate back to the originals underlying them.
Specifically for papers the best thing we have is probably MeSH. That may not be perfect but it is also under used. A lot of papers that are in Pubmed are annotated with MeSH terms. It would often be better to use these instead of trying yet another annotation or mining approach.
Thank you for your insights. The best thing I had in mind so far concerning a larger term database was indeed the MeSH DB. Concerning a "perfect" ontology. i.e. in the sense of a non-overlapping one, this is of course unrealistic, because e.g. biological structures interact with their environment in a nonlinear fashion and thus their properties may evoke different terminologies through different viewpoints. Not to speak of evolution.
Still, there should be an easily accessible search, answering queries in the form of a dictionary-esque phrase with references and not in the form of a db nodemap.
Example query: (What is a) "driver gene" (?)
Answer is given implicitly in some paper (not dictionary-ready):
Cancer is a disease of the genome. As cancer initiates and progresses, genomic instability develops and abnormal genomic changes (e.g., sequence mutations, aberrant promoter methylation, and structural lesions such as gains/losses, inversions and translocations) accumulate. While some of these abnormalities disrupt normal cellular processes and contribute to cancer initiation and progression (i.e., drivers), others emerge simply as victims of genomic instability occurring as a result of cancer progression (i.e., passengers). doi:10.1186/1471-2407-10-426
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6