Introduction

A causal relationship is defined by a correlation between two events or elements where one influences the other. In Molecular Biology, causality underlies the principles of interactions between biological components, for instance in governing regulatory mechanisms. Consequently, the understanding of these interactions and the mechanisms that they enable, involved in biological regulations (e.g., gene regulation) is to be improved by looking at the core interactions happening among entities. Once causality is established, the next challenge is to adequately represent and archive it so that it can be shared and used by computers and humans alike. This repository is intended for:

• providing a state of the art of existing resources and formal representations depicting causal relationships among biological components
• developing and documenting mappings between resources/languages capturing causal reasoning information in Molecular and Cellular biology.

Existing Resources depicting causal relationships:

• The Gene Ontology - schema_docs
• GRECO - schema_docs
• Signor - schema_docs
• SignaLink - schema_docs
• InTact - schema_docs
• Causal Biological Network - schema_docs
• Reactome - schema_docs

Languages:

• OpenBel
  • Open data in OpenBel:
  • RDF conversion available
• BioPax
• SBMLqual
• CausalTAB

Tools:

• CausalR - schema
  • Open data in CausalR:

Each of these resources captures, or is begining to capture causal information. There are many commonalities between the causal information being captured: multiple resources record whether causal relations are -ve vs +ve, direct or indirect & record mechanism. This site aims to document and align the vocabularies and semantics used by each of these resources.

Strategy for sharing data:

We recognise that each resource already has a representation that is doing useful work, so we seek to map existing schemas rather than propose changes. To do this we will:

• Map all entities to standard ontology classes/relations.
• Map statements to a language with defined semantics: RDF triples? OWL? ...

Differences between resources:

• IntAct, Signor, CausalR: causal & regulatory relations apply between gene products. Signor records the processes involved separately as 'mechanism', maintaining the association between regulation and process within each statement.
• GO: causal & regulatory relations apply between the processes that gene products are capable of. This has some advantages. A gene product may have multiple functions - only one of which is a mechanism for regulating the activity of a second gene product.

One obvious way to bridge this gap is for RO can define shortcut relations mapping to the GP:GP relations in other resources, something like this:

• name: has_activity_that_positively_regulates_activity_of
• xref: signor:activates
• xref: causalR:activates
• Property chain: capable_of o positively_regulates o enables -> has_activity_that_regulates_activity_of

Where mechanism is recorded, we could potentially use a simpler relation:

GP1 activate GP2 mechanism:phosphorylation => GP1 capable_of some (phosphorylation that regulates_activity_of GP2)

Proposed structure for mapping files:

Relations table (signor example is fictional - for illustration purposes only)

resource name	relation name	relation ID	domain	range	description	example usage
RO	regulates	RO_	process	process
signor	activates	-	protein	protein

Class table

resource name	class name	class ID	domain	range	description	example usage
PRO	protein	PRO_	process	process

Mapping table

resource name	relation name	relation ID	ontology relation name	ontology relation ID

Repository structure:

  doc/
    README.md
     GO/
        README.md
     GRECO/
        README.md
     Signor/
        README.md
     SignaLink/
       README.md
     IntAct/
       README.md
     CBN/
       README.md
     Reactome/
       README.md