Mechanism of Action

For any stimulus producing a specific response in a biological system, systems biology approaches may predict its molecular mechanism, or mechanism of action (MoA). (In a strict sense, mechanism of action normally refers to molecular mechanisms of drug action, but we also use it in a broader sense referring to any molecular mechanism, such as the ones deriving from protein/gene activations or inhibitions.)

Even when the physiological effect of a stimulus is exactly the same for two individuals, its molecular mechanisms may be different. The same applies to biological systems other than whole organisms (organs, tissues, cells…): indeed, probably there are many pathways connecting any given stimulus with its response. However, they will probably share common patterns. Thus, the global MoA will contain different, related pathways which the stimulus will modify variedly in different individuals.

With all this in view, there are some questions to be solved:

Is it possible to draw a MoA?

Not really. Each MoA developed by SIMScells for a stimulus–response pair is generated from more than 20,000 individuals or mathematical solutions accurate enough to be considered possible. The whole MoA would be therefore extremely difficult to draw in a single picture.

Then, what is really a MoA?

SIMScells' MoA pictures represent the average of all the mechanisms calculated for a stimulus–response pair. If the variance of the total set of mathematical solutions is not very high, this average will be a good representation of all the MoAs.

What does the MoA picture shown in the SIMScells' report mean?

SIMScells' MoA pictures represent the average of all the mechanisms calculated for a stimulus–response pair. If the variance of the total set of mathematical solutions is not very high, this average will be a good representation of all the MoAs.

How to interpret a MoA representation?

  • Protein activities are encircled as nodes
  • The effect of a node (protein activity) over another is represented as a link between them:
    • · When the activity of the first node activates or enhances the second, the link is a green arrow.
    • · When the activity of the first node inhibits or inactivates the second, the link is a red line.
    • · When there is a known interaction, but the effect of the first node over the second cannot be specified as "activation" or "inhibition", the link is a blue line finishing with a diamond.

For example, in this MoA representation, the activity of the stimulus inhibits or downregulates THRB, whereas this protein, when activated, inhibits GPIBA. Knowing that, the MoA may be read as an expansion of the stimulus: the stimulus inhibits THRB, promoting GPIBA activation, what induces the activation of VWF, and thus precludes ITA2 activity. At the end, the response is activated.

Validation of the mechanism of action

Statistic parameters are calculated for each MoA analysis in order to have a way to validate and assess the quality of the model. These parameters are dependent on, among other variables, the time of calculation and the available scientific knowledge about the components of the MoA.

  • Accuracy: It is the percentage of fulfilment of the restrictions, the known information used to construct the mathematical model. In other words, how much the models comply with the scientific data used to train it.
  • Coverage: Refers to which percentage of individuals of a population are included in MoA drawing, i.e., how many individuals of the population are represented in a specific cluster. This means that the coverage of the Global MoA Representation will be always 100%, since this MoA Representation is constructed from all the population. The coverage is a measure of the relevance of the clusters within the virtual population: more coverage means that more individuals of the populations are contained in that cluster. Usually, the clusters with higher coverage are also the simpler in statistical terms. However, it has to be taken into account that sometimes, in Nature, the simplest solution is not always the best, so all clusters have to be analysed although having low coverage values.
  • Uncertainty: This is an indicator of the variance of the different MoAs solutions obtained. More uncertainty means more complexity and more possibilities of explaining the Truth Table. Thus, it is an indicator of the complexity of the MoA. When MoA solutions are very similar, they are easier to average in a unique MoA representation, what would mean a low uncertainty. On the other hand, when uncertainty is high, meaning that the calculated MoAs solutions are very dissimilar and difficult to average, it is necessary to analyse the results from the individual MoA clusters to obtain more accurate information.

What kind of information is reported in the SIMScells' MoA Reports?

Different MoA products are offered by SIMScells to perfectly fit your research needs. More information can be found in the MoA products section and examples of MoA Reports are available to download.