Truth Table

In order to simulate in silico the behaviour of a biological system, we have to train its virtual model to respond to specific stimuli like the real system does. This requires feeding the protein network with stimulus–response relationships exhibited by the real system.

The Truth Table is a database containing such stimulus–response pairs (restrictions), manually extracted from the scientific literature by SIMScells' expert team, that the virtual system must obey. It is specific for the definition of each virtual biological system (see Virtual BioSystem Definition).

The Truth Table includes three types of data:

  • a) Clinical information about drug targets and known effectors: The treatment of any medical patient with a chemical compound may be considered as an in vivo experiment. Thus, the clinical databases listing drug pharmacological effects (DrugBank, SIDER…) provide us with a large amount of precious information about the behaviour of the human organism. SIMScells uses these databases to extract information about drug targets, pharmacological action, off-targets, adverse events, indications and mechanisms of action. Since training the model requires a language which it understands, drugs and biological effects included in the Truth Table are translated into drug targets and biological effectors through the Biological Effectors DB (AX-Health DB). For example, we know that aspirin relieves headache, but our mathematical model, built with proteins/genes as nodes, does not understand what aspirin or headache are; it is necessary to translate it to molecular level.
  • b) Information from microarrays: The Truth Table contains curated information from high throughput experiments generated by different technologies and available from public resources (GEO database).
  • c) Mutations and effectors: The Truth Table also includes information about the effects of gene knockdowns and mutations. The affected gene(s) or protein(s) are associated to their biological responses at a biochemical or clinical level.

SIMScells offers its users the possibility to customize the training of their model by loading their own information about it: microarrays, knock-out information, novel interactions discovered …