Different types of molecular-interaction databases

There are a large number of publicly available molecular interaction databases. These can be classified into three main types, according to their data-acquisition policies 14.

Primary databases are those that collect experimental molecular interaction data exclusively from peer-reviewed scientific publications. IntAct 11, MINT 10 and MatrixDB 16 are examples of this type of database. They can be further classified by the level of detail that they use to represent the information and the depth of their curation policies.

Secondary databases, also known as meta-databases, seek to integrate the data curated by several primary databases in one, integrated repository. APID 17 and PINA 18 are examples of this type.

Finally, predictive databases combine the experimentally inferred data taken from primary databases with computational predictions of molecular interactions. Examples include STRING 19 and UniHI 20.


Now, let's have a look at the different types of curation that primary databases use to register molecular interaction data.