ModRev - Model Revision Tool for Boolean Logical Models of Biological Regulatory Networks

Abstract

Biological regulatory networks can be represented by computational models, which allow the study and the analysis of biological behaviours, therefore providing a better understanding of a given biological process. However, as new information is acquired, biological models may need to be revised, in order to also account for this new information. Here, we present a model revision tool, capable of repairing inconsistent Boolean biological models. Moreover, the tool is able to confront the models, both with steady state observations, as well as time-series data, considering both synchronous and asynchronous update schemes. The tool was tested with a well-known biological model that was corrupted with different random changes. The presented tool was able to successfully repair the majority of the corrupted models.

This work was supported by national funds through Fundação para a Ciência e a Tecnologia (FCT) with reference SFRH/BD/130253/2017 (PhD grant) and UIDB/50021/2020 (INESC-ID multi-annual funding).

A Tutorial

The model shown in Fig. 1 is represented by the following listing:

Now let us consider that we want to define a steady state observation in which $v_1$ has value 0, $v_2$ has value 0, and $v_3$ has value 1, as following:

Using ModRev tool, giving the model defined above (as a file model.lp) and the steady state (as a file obsSS.lp), execute the following command:

This output means that the model in Fig. 1 can be repaired by changing the interaction type between $v_1$ and $v_3$. If we repair the model and execute the above command again, the result will be:

Now let us assume that the user knows that the interaction between $v_1$ and $v_3$ is correct, and wants to prevent repairs over it. The predicate fixed(v1,v3). can be used to define that the edge between these nodes can not be changed or removed. Adding this predicate to the model and running the command above, we obtain the following result:

A different set of repair operations is obtained that does not change the fixed edge. Now assume that the user wants to prevent any repair over the node $v_3$. The predicate fixed(v3). can be used to prevent that node to be inconsistent. However, in this example, if we prevent any change to node $v_3$, considering its regulatory function, and that $v_1$ has value 0 and $v_3$ has value 1, and we are in the presence of a steady state, it becomes impossible to repair the network. In this case, when the model is over-constrained, using the same command as before, the tool produces the following message:

Consider now that we have, for the same model in Fig. 1, a time-series data as shown in Table 2. Consider that this experimental observation with three time-steps (0, 1 and 2) is considering a synchronous update scheme.

Table 2. Synchronous time-series data

We can represent the time-series data using the following listing:

Note that we start the file indicating the maximum value of time step with #const t = 2.

Using ModRev to verify whether the model is consistent, while considering the above time-series data (as a file obsTS01.lp) under a synchronous update scheme, execute the following command:

This will produce the following result:

Note that now we have multiple choices to render the model consistent. To repair node $v_2$, for example, one can apply the operations in Repair #1 or in Repair #2. The same applies to repair node $v_3$.

If instead of a time-series data under a synchronous update scheme, we are under an asynchronous update scheme, the previous command would change from -up s to -up a. The option -up indicates the update scheme to be considered, with argument s for synchronous and a for asynchronous.

ModRev also supports incomplete time-series data. Assume that we have the experimental observation shown in Table 3, where node $v_3$ was not observed, and a value of $v_1$ was also not observed.

Table 3. Incomplete synchronous time-series data

Consider the following representation of an incomplete time-series data:

Executing the following command, while considering the above experimental observation (as a file obsTS02.lp) under synchronous update scheme, produces the result below.

ModRev tool also supports confronting a model with multiple experimental observations at the same time. For example, we could confront the model of Fig. 1 with the two time-series data above, using the command:

Note that the directive #const t = 2 must only be defined once.