Page tree
Skip to end of metadata
Go to start of metadata

Goal

  • Represent relationships where the source (cause) and target (effect) are specified, but the direction of causal influence (increases/decreases) is not.
    • These relationships are identified in text mining and database conversion
    • In BEL v1.0, these are translated to A association B, resulting in loss of information

Proposed Representation

relationship type = causal

The regulates relationship should be used when A is reported to affect B, but it can not be determined if A increases or decreases B. For example, if a paper states that change in the quantity of A results in a decrease of B, it is not clear if an increase in A or a decrease in A is responsible for the decrease in B. This could be represented as:

A regulates B

This relationship provides more information than association, because the upstream entity (source term) and downstream entity (target term) can be assigned.

Proposed default compiler handling

If a regulates edge connects the same source and target as an increases or decreases edge (including directlyIncreases and directlyDecreases), the relationship providing the most information should be used. For example, the following statements should compile to single “increases” edge, associated with both statements:

A regulates B

A -> B

Proposed application use

For scoring applications like Whistle, causes edges should be treated as ambiguities, i.e., similar to source/target pairs connected by both causal increase and causal decrease relationships. For Whistle, the ambiguous targets are counted towards richness, but not towards concordance.

  • No labels