Accessing biological data as Prolog facts

Abstract

It has been argued before that Prolog is a strong candidate for research and code development in bioinformatics and computational biology. This position has been based on both the intrinsic strengths of Prolog and recent advances in its technologies. Here we strengthen the case for the deployment and penetration of Prolog into bioinformatics, by introducing bio_db, a comprehensive and extensible system for working with biological data. Our library packages high quality, publicly available biological databases that are routinely used in tasks such as: (a) the translation between biological products and (b) product-to-product interactions which can be visualised as graphs. This library allows easy access to these data in five formats: Prolog fact files, Prolog quick load files, Berkeley DB data files, RocksDB and SQLite databases. In addition, the library introduces two innovative features that are pertinent to data analytics in general. First, on-demand downloading of prepacked data files as well as reconstruction from latest data files from the curated databases are supported. Second, by employing code hot-swapping the library delivers the data: (a) transparently to the user and (b) in the familiar format of Prolog facts.