FADB-China Implementation

FADB-China was developed on the basis of the Linux, Apache, MySQL, Python (LAMP) strategy; the server runs under apache2 on a Linux machine running Ubuntu Server. The algorithm and backend program were written in Python by using the Django framework in combination with MySQL to manage the data. Hypertext Markup Language (HTML), Cascading Style Sheets (CSS), and JavaScript were used to implement FADB-China’s front-end data presentation and interactions.

Query Methods

FADB-China developed multiple retrieval methods based on cheminformatics algorithm, such as structure retrieval, fragment retrieval, similarity retrieval, maximum common substructure (MCS) retrieval, and text retrieval options. JSME structure editor is included in the search option to enable searching through the drawing and editing of molecules.

Data Sources

To obtain a comprehensive record of Chinese food adulteration, we systematically compiled more than 500 related articles, national standards, and announcement documents issued by the Chinese government. Then, the food adulteration data were manually extracted from the literature and regulatory standards, and the data were detail-annotated with the illegal additive substance, main effective components, addition purpose, involved links, related foods, and reported detection methods.

Illegal Additives Prediction

On the basis of previous research, drugs, pesticides, veterinary drugs, industrial dyes, and industrial chemicals are often added to food in violation of regulations because they are easily obtained. We selected 26 related compound databases and lists, which are managed by authoritative international organizations or published in peer-reviewed journals.

We obtained the compound name, its chemical structure, database indexes, and original use from these compound databases and lists; these were then added to the list of potential illegal food additives. We normalized the compounds’ SMILES in the candidate list and used this to combine the same molecules from different data sources into one record.

We chose Daylight fingerprints and extended-connectivity fingerprints (ECFPs) to describe the molecular structure of illegal additives.

We chose the Tanimoto coefficient-based similarity algorithm and fMCS algorithm to calculate the molecular similarity. We used 0.85 as the threshold for screening, which is generally considered to have similar activities; the result of any algorithm that exceeds the threshold is considered effective, and higher similarity implies a higher likelihood that they will have similar activities or functions.

Release Notes

V 1.0 2019.09.10 The first version of the FADB-China database has been released.

V 1.1 2019.10.20 Optimized image quality in database.

V 1.2 2019.11.24 Added visitor statistics function.

V 1.3 2019.11.30 Optimized database introduction and FAQs.

V 1.4 2020.01.10 Added 16 food adulteration cases uploaded by scientists.