In silico media & subsets database

This section describes the architecture of the database, as well as information on what and how the data is stored.

Entity-Relationship-Diagram

Entity-Relationship-diagram for the database with exemplary tables.

The diagram above shows the Entity-Relationship-diagram (ER-diagram) for the database, as well as examples of the tables that the final database consists of. The ER-diagram displays the four entities medium, substance, subset and database (blue), their relations (orange) and attributes (green). Each medium includes multiple substances from different biological origins that can be absorbed by the model with the stated fluxes. Similarly, each subset describes by name and a description, includes different substances with specific percentages. Since multiple media and subsets can include the same substance, the relationship is marked as n:m. Each substance is described by a name and a formula in addition to an unique identifier. Furthermore, the substances can be found in different databases and connected to the respective identifiers. Since this 1:n relation does not include further attributes, the tables were combined, forming a total of six tables in which the data is stored.

Medium

The medium table contains an entry for each medium in the database, giving it an unique ID and storing its name, description and reference. The name should be a short abbreviation that is generally used to refer to the medium, e.g. LB. The description should be a short description explaining its abbreviation, e.g. for LB the description “lysogengy broth” explains the abbreviation used for the name. The reference is the DOI or link to the original in vitro definition of the medium or an already existing in silico definition.

Medium2substance

This table links the medium ID to the substances contained in a medium. Additionally, the flux for the possible exchange reaction is stored. In case the original publication of the medium did not provide information on the flux, the attribute is left empty (NULL). Furthermore, since the chemical substance in the medium might differ from the compund used to create the medium, the column source provides the information about how and in which form the substance was originally added to the medium.

Subset

The subset table contains an entry for each subset in the database, giving it an unique ID and storing its name and description. The name should be a short abbreviation of the subset’s full name, e.g. artSe for ‘Artificial Sebum’. The description should be a short description explaining its abbreviation and if possible containing a link or doi referencing the original composition, e.g. for artSe the description “Artificial Sebum (based on: https://journals.asm.org/doi/10.1128/spectrum.04180-22, Table 2)” explains the abbreviation used for the name and provides a reference.

Subset2substance

This table links the subset ID to the substances contained in a subset. Additionally, the percentage for the possible exchange reaction is stored. In case the original publication of the subset did not provide information on the flux, the attribute is set to 1.0 for 100%.

Substance

This table consists of the substance ID, its name and its chemical formula. The chemical formula is saved without charge. In case of substances like vitamin B12, no substance formula is saved as these substances describe a group of different metabolites, that are treated as a single substance in some reactions.

Substaces are added using the following naming conventions:

  1. The names are mostly based on ChEBI.

  2. For amino acids, hormones and the like, the trivial name or otherwise one of the IUPAC names on the ChEBI entry with the most used alias in square brackets is used.

  3. For acids, their salt form, e.g., sulfate for sulfuric acid is utilised.

  4. Metallic ions are entered using the following scheme:
    Name(charge in roman numerals) cat/anion [chemical formula(charge in arabic numerals)]
    e.g. Calcium(II) cation [Ca(2+)].
  5. Halogenic ions are saved in a similar matter,
    e.g. Chloride [Cl(-)].

Note

When coverting the substances from the previous database format to this, well-annotated names were taken from the previous table as they were - meaning that they do not necessarily follow the rules defined above. These names might be subject to future change. Any newly added substance should follow the naming scheme described above. For more information about the conversion, see the jupyter notebook Remapping_substance_table.ipynb

Hint

Before adding a new substance to the table, please check carefully that the substance does not already have an entry.
In case your substance already exists, please use the existing entry.
Help regarding mapping your information e.g. the BiGG identifier to a formula or the ChEBI IUPAC name, please refer to the functions XXX and YYY.

Substance2db

The final table of the database contains the mapping of the substance identifiers to their respective various database identifiers. To ensure no problems occur in the case of multiple databases with the same identifiers and to help with searching for a certain kind of ID, the database type is stored in an additional column. Currently, IDs from the following databases can be found in this table:

Hint

If the ID of two databases is the same (this can occur with e.g. BiGG and VMH), the database type is comprised of both database names, separated by a plus sign (e.g. ‘BiGG+VMH’)

Note

Be aware that it is possible for some substances to have missing IDs for certain databases in the list above as they either do not exist or have yet to be entered into the database.