With recent changes in the policy of the National Science Foundation (NSF) including the assessment of scientists not only by all the papers they published, but by all ‘research products’ they make available to the scientific community , the quest has begun for new ways of sharing scientific data. One way of sharing scientific information beyond the peer-reviewed journal system is publishing scientific findings in ‘nanopublications’.
Basically nanopublications can be regarded as ‘minimal publications’, containing only the two most essential parts of a publication: information about the author(s), and scientific information.
Like any other publication, nanopublications serve the purpose of sharing and spreading scientific information. Generally, the way scientific information is gathered has changed dramatically, especially in the Life Sciences: Experiments are no longer exclusively done in order to verify or falsify a hypothesis. Large-scale screens simply collect information without a prior hypothesis being necessary. The same is true with negative data: they do not fit the prevalent hypothesis, thus, the true hypothesis which these results support is potentially unknown. Classical publishing prefers research in which the data and the hypothesis are congruent. There is a growing interest, however, to publish scientific findings which do not directly support a specific hypothesis, for two reasons: First, this kind of data is produced in large amounts and should be communicated to make best use of it, and second, these findings are valuable ‘research products’ which can give credits to the scientists who made them.
Although there are no binding standards for nanopublications, there are some efforts under way to create standards for the format of nanopublications , . These standards should serve the purpose of making nanopublications easily accessible in a unified format, i.e. for automated reasoning of infrastructures. Emerging standards employ the Resource Description Framework (RDF)-data model . Briefly, the RDF data model uses a so called triple to make assertions of resources in a subject-predicate-object manner. For example: air(subject) contains(predicate) oxigen(object) would be an easy example of a semantic triple used in RDF. Current standards, i.e. as being suggested by the Open PHACTS consortium , divide the communicated content of a nanopublications into
Both the assertion and the supporting data should be structured in hierarchies of RDF-triples.
Principal structure of a nanopublication. Modified from: http://nanopub.org/original
Since there are no binding standards for the format of nanopublications, we have not found any dedicated platform for their distribution. Nevertheless, the concept of nanopublications is very interesting, given that there is or will be a platform on which you can publish scientific data in a standardized format.
We are excitedly watching the development of open standards and emerging nanopublication platforms. The idea of nanopublications to make the exchange of scientific data readily available – which is not necessarily referring to a hypothesis, together with the idea to make these pieces of knowledge machine-readable – highly correspond to the philosophy of labfolder.
Given the fact that labfolder contains structured data sets of empirical knowledge and methodology, we hope that very soon, we can make significant contributions to the development, the reach and the impact of nanopublications.