Scientists from the National Institute of Standards and Technology (NIST) in the USA have developed networks of molecular data that, like Facebook, can be used for data and graph-based searches. Scan a huge amount of chemical and biological data in order to find substances with desired properties. For example - substances with five rings that have an affinity for a certain enzyme
The social network expanded and reached a new and surprising destination: molecules. Scientists from the National Institute of Standards and Technology (NIST) in the USA have developed networks of molecular data that, like Facebook, can be used for data and graph-based searches. While a graphic search on Facebook will allow users to find all their colleagues in New York City, for example, who like to drink beer through a quick search, the networks developed by the scientists will be able to help researchers scan a huge amount of chemical and biological data in order to find substances with desired properties. For example - substances with five rings that have an affinity for enzyme A. This search approach could accelerate the development of new drugs and other desirable substances.
Molecules do not have an individual profile on the web, similar to Facebook users, so the researchers' first challenge was to develop a common language for scientists to use to describe the objects of their research. For example, one research group may describe the properties of the material as "vitreous", while another group may actually use the definition "vitreous", even though both definitions have the same meaning, explains one of the research partners.
One of the possible approaches to solving this issue could be to define a set of standard terms, but the researchers preferred a more flexible approach that could evolve over time. The search language that the researchers devised is similar to Indo-European languages, such as Sanskrit and Latin, in which short roots are used to build words while relying on a set of rules. One of the researchers gives as an example the word "yoga" in Sanskrit based on the roots "y" which means friend, "v" which means the creator (God) or the mind, and "ge", which means movement. Similarly, scientists could take three simple root words "red", "laser" and "light" and combine them to describe a word for a definite compound called "light-laser-red" that expresses a new idea. The use of the root and the rules of the language means that scientists who know the root words will be able to understand the meaning of unfamiliar terms. In addition, this language allows scientists the flexibility to easily develop new non-standard terms in the future.
The team of researchers has already applied its new dictionary rules to the existing chemical structures in the scientific database called PubChem, a massive database of millions of compounds and chemical substances, in the World Protein Database (PDB) and in databases used by the institute itself. Although the scientific databases have not yet reached the volume of Facebook's activity of more than a billion users, they are actively used by many scientists inside and outside the institute.
From the moment the initial dictionary was established in this language, the researchers also worked to label the descriptions of the molecules and scientific experiments in a scalable way that would allow scientists to reach both comprehensive and accurate results. A well-known problem with many search methods is getting too many results, notes the lead researcher. The researcher describes the approach of his research team as similar to the challenge of finding a bag of potato chips in a large department store - first you find the appropriate store, in the second stage of the grading you find the snack area, there is the fried area and from there you will soon know if there are any potatoes at all "Yeps or not," explains the lead researcher. "That way, even if the store has millions of items, you can tell pretty quickly if they sell potato chips or not." The team of researchers claims that their ranking method can also help scientists who need to find keywords to use in their research articles.
Organizing the large amounts of information produced in the field of science is a great challenge, explains the researcher, but these amounts may have many benefits. Effective graphical searches could allow scientists to quickly locate chemical structures and physical properties they require in order to develop new drugs or advanced materials.
