loucerac/dmaps

Placeholder for common development on hipathia and disease maps.

Hipathia and Disease Maps

Development repository to process the Disease Maps map with Hipathia.

Install

• Clone the repository
• Create a .ENV file
  • Copy example.env to .env -- Change the maps_dir environment variable as outlined in example.env.
  • Load the project using RStudio -- If the requirements are met, packrat does everything.

Requirements

• R >= 3.5
• The SBML library must be installed and reachable by R.

How to contribute

• Clone the repository.
• Initialize git-flow.
• Use the develop branch.
• For the moment, create a notebook following the naming scheme of those already included.
• Once we begin to produce real code, follow the standard git-flow work-flow.

Usage

At the moment only exploratory notebooks are included.

Graph parsing

Notebooks 1 and 2 represent different experiments of the same approach: parse the SBML map with the read_bml library, convert to igraph, using the graph representation included in the library, and then work with the induced attributes.

The automatic graph conversion provided by the read_sbml library poses serious problems: for instance, the edges pointing to other edges, such as those labeled as modifiers, are missing. In the end most inhibitions are lost during the conversion.

Metadata parsing

• Reactions labeled as HETERODIMER_ASSOCIATION can be removed as they represent complex-formation processes.
• For each reaction that includes modifiers we include one auxiliary node (with input value 1) named aux.
  • Each node labeled as reactant is connected to aux as input and output, respectively.
  • Each node labeled as product is connected to aux as output and input, respectively.
  • Each node labeled as modifier is connected to aux as input and output, respectively.

Figure 0 shows he most basic example, with one product, one reactant and two modifiers.

Figure 0: Basic modifier example.

For example, suppose that we have the following sequence s1,s2/s5,s6/s3,s4 where the nodes are labeled as reactant/product/modifiers. Figure 1 shows the resulting network after we apply the previous rules.

Note, however, that we have provided the most general example, as the number of reactants, products and modifiers can be any number greater than 0.

Figure 1: Modifier network reconstructions.

TODO

• Resolve BOOLEAN_LOGIC_GATE_AND

Figure 2: boolean_logic_gate_and.

• Resolve BOOLEAN_LOGIC_GATE_OR

• Resolve those nodes labeled as modifier that are parsed as CATALYSIS and are not properly resolved because there is another tag, either BOOLEAN_LOGIC_GATE_AND or BOOLEAN_LOGIC_GATE_OR, which has not been correctly parsed.

• Resolve protein_X to protein_X

Figure 3: Same protein interactions.

• Resolve modulations

Figure 4: Modulations re1

Figure 5: Modulations re3

• Include the module parsing functionality from the graph notebooks.