LEGO Specification Proposal

Huaiyu Mi and Paul Thomas

Oct. 10, 2014

Annoton

Definition: An annoton is an annotation unit to describe a biological activity in context of another biological activity that it regulates. It can be simply illustrated as two biological activity units and the effect relation between them (Figure 1). This is the smallest information unit in LEGO annotation.

caption

Figure 1. The basic Annoton unit.

Biological activity

Definition: Biological activity is defined as activities and processes carried out by all biological entities. Specifically, it is the molecular function that one or more biological entities are capable of performing, via a particular biochemical mechanism, in a specific cellular location, as part of a biological process. Please note that the function here is just the “potential”. Some activities (those under “molecular activity regulator” in the overhauled MF ontology) do not constitute executable function without the target function and the molecular effect relation (see below).

The biological activity unit is composed of the following classes of information (Figure 2): active entity, molecular function, cellular component, biological process, and context relation. Below is a description of each of these classes.

caption

Figure 2. Classes of information in biological activity.

Active entity

Definition: A biological entity that is in an active conformation with enough quantity to carry out the molecular function. This is usually the gene(s), gene product(s), or a protein complex that is annotated in GO.

Usage: The active entity can be 0 (when it is unknown or unspecified), one or more than one. When more than one entity is assigned, the following logic operators should be used to indicate their relationship:

• AND: all entities are required for the activity;
• OR: either entity is required for the activity;
• NOR: neither entity is required for the activity.

Molecular function

Definition: This is the GO molecular function that is carried out by the active entity. Usages:

• Each biological activity can have 0 (when root MF term is assigned) or one molecular function term assigned to it. This can be the MF annotation to the gene(s) or gene product(s) in GO annotation.
• When the root MF term is assigned, it means the function is unknown or unspecified. This can be used to indicate a biological process in general while specific function is not specified. For example, xxxx.
• When a gene or gene product is annotated by more than one GO molecular function terms, they have to be represented by multiple biological activity units.

Cellular component

Definition: The location of the active entity when it executes the molecular function. This is the GO cellular component annotated to the gene or gene product.

Usage:

• Each biological activity unit can have 0 (when root CC term is assigned) or one GO cellular component. This can be the CC annotation to the gene(s) and gene product(s) in GO.
• When root CC term is assigned, it means the cellular component is unknown or unspecified.
• No more than one CC term can be assigned to a particular biological activity unit.

Biological process

Definition: The biological process that the molecular function and the active entity are involved in or part of. This is usually the BP annotation to the gene(s) and gene product(s) in GO.

Usage:

• Each biological activity unit can have 0, one or more than one GO biological process (if unknown, the root BP term is used).
• If there is more than one, the processes must be related to each other via part_of or is_a relations.
• The special case of biological process is outcome, in which case the biological activity unit on a biological process that the molecular activity (or active entity and molecular function) is not a part of. It is used when an experiment measures a larger process (or some marker for that process) and not a molecular activity that is part of a process (in which case the molecular effect relation should be used).

Context relation

Definition: These are the relations that describe the relationships of different classes within the biological activities as well as the relation of the unit to external ontologies such as Cell Type, Development Stage and Tissue Ontology. For example, a molecular function is located_in a cellular component that belongs to a particular cell type.

Effect relation

Effect relation defines the relationships between the two biological activities units. There are two subtypes of effect relation: molecular effect relation and process effect relation.

caption

Figure 3. Effect relation has type subtypes: molecular effect relation and process effect relation.

Molecular Effect relation

Definition: The molecular effect relation describes the effects exerted by one biological activity unit to a molecular activity unit. It mainly represents the same semantics as the “regulation of molecular function” classes currently in GO. It includes the following relations.

• Directly activate
  • Definition: A direct action, usually involving direct protein-protein interaction, that produces a positive or activating effect from one molecular activity unit to the other.

• Directly inhibit
  • Definition: A direct action, usually involving direct protein-protein interaction, that produces a negative or inhibiting effect from one molecular activity unit to the other.

• Positively regulate
  • Definition: An indirect action that produces a positive or activating effect from one molecular activity unit to the other.

• Negatively regulate
  • Definition: An indirect action that produces a negative or inhibiting effect from one molecular activity unit to the other.

• Regulate
  • Definition: An action, either direct or indirect, that produces a regulatory effect from one molecular activity unit to the other. This relation does not specify whether the effect is positive or negative. It is used when the effect is either positive or negative.

• Upstream of
  • Definition: This relation indicates that one molecular activity precedes the other. It has no description about any regulatory action from one to the other. It can be used when the regulatory effect is unknown or irrelevant in the model.

• Required for
  • Definition: This relation indicates that the activity is required in order for another activity to occur, but it is constitutive and not regulatory.

Process effect relation

Definition: The process effect relation describes the effects exerted by a biological activity unit on a biological process. There are two situations when a process effect relation should be used.

1. The effect of a biological activity unit on a biological process that the active entity and molecular function are not a part of. It is used when an experiment measures a larger process (or some marker for that process) and not a molecular activity that is part of a process (in which case the molecular effect relation should be used).
2. The effects of a biological activity unit on a molecular function and a biological process are the same. For example, IKK directly activated IkB (an NFkappaB inhibitor), and thus positively regulated NFkappaB signaling.

It includes the following relations. It mainly represents the same semantics as “regulation of biological process” classes in the GO, which are not themselves part of LEGO.

• Positively regulate
  • Definition: An indirect action that produces a positive or activating effect of a biological activity unit on a biological process.

• Negatively regulate
  • Definition: An indirect action that produces a negative or inhibiting effect of a biological activity unit on a biological process.

• Regulate
  • 'Definition: An action, either direct or indirect, that produces a regulatory effect of one biological activity unit on a biological process. This relation does not specify whether the effect is positive or negative. It is used when the effect is either positive or negative.

• Upstream of
  • Definition: This relation indicates that a biological activity precedes a biological process. It has no description about any regulatory action from one to the other. It can be used when the regulatory effect is unknown or irrelevant in the model.

• Required for
  • Definition: This relation indicates that the activity is required in order for a process to occur, but it is constitutive and not regulatory.

• Combining effect relations

Effect relations from more than one biological activity unit can be combined using logical operators:

• • AND: The effect occurs only if all combined activities are executed.
  • OR: The effect occurs if at least one of the combined activities is executed.