go-cam-user-guide

name	go-cam-user-guide
description	Use when starting a completely new GO-CAM model from scratch, deciding whether a GO-CAM is needed vs standard annotations, or following the official GO Consortium workflow for model creation. Covers the conceptual structure of GO-CAMs, when to use them, and the step-by-step creation workflow.

GO-CAM annotation guidelines

Standard GO annotation versus GO-CAM 1

Aim of GO-CAMs 1

When to make a GO-CAM versus a standard annotation 1

Suggested Workflow to create a GO-CAM model 2

GO scope with respect to activity mechanism 3

Model Title 3

Review status 4

Revision history 4

Standard GO annotation versus GO-CAM {#standard-go-annotation-versus-go-cam}

Please refer to the GO website for a description of standard GO annotations and GO-CAMs.

Aim of GO-CAMs {#aim-of-go-cams}

GO-CAMs (Gene Ontology Causal Activity Models) are designed to represent:

1. ‘activity units’ that specify how different GO annotations (MF, BP, CC) for a given gene product relate to each other.
2. ‘pathways’ that specify the flow of causal connections between the activities of different gene products, to accomplish or regulate a larger biological process.

Formally, GO-CAMs are constructed by joining together standard GO annotations (one gene product and one GO term) into a larger model of a biological process or pathway. In this way, standard annotations can be used to help build a GO-CAM model, or, conversely, a GO-CAM model can be decomposed into standard annotations.

When to make a GO-CAM versus a standard annotation {#when-to-make-a-go-cam-versus-a-standard-annotation}

GO-CAM models are valuable to depict a biological pathway or process in which multiple activities can be interconnected into a causal flow. However in cases where activities cannot be connected, either because they have no causal relations, or the causal relations are unknown, then one or more standard GO annotation(s) will suffice to describe the data.

Figure: GO-CAM structure. The dotted box shows an activity unit. Multiple activity units are joined together by causal relations that connect the ‘Molecular activity’ part of the activity units (in the diagram, the other parts of the second activity unit are omitted for simplicity). Evidence is associated with each edge (arrow) in the model, so that complex models can be built up from simpler pieces of knowledge obtained from different experimental systems.

Suggested Workflow to create a GO-CAM model {#suggested-workflow-to-create-a-go-cam-model}

It can be very helpful to refer to a summary figure from a recent research article or review to help visualize a potential GO-CAM.

When making a biological process-centric GO-CAM model, we suggest these steps:

• What are the main activities (MFs) of each of the gene products in a model?
• How do those activities relate, in a causal chain, to each other?
• What processes are those activities involved in?
• Where do the activities occur?

GO scope with respect to activity mechanism {#go-scope-with-respect-to-activity-mechanism}

Binding terms and mechanistic descriptions are out of scope for GO-CAM models.

Avoid using Molecular Function binding terms as they do not adequately describe the protein's activity. Non-catalytic MF can be harder to identify, see documentation on non-catalytic MF for suggestions to replace 'binding' terms.

Mechanistic description of the function belongs in the definition of the MF term rather than in the GO-CAM model. For example, a tyrosine protein kinase receptor’s activity could in principle be broken down into separate steps like ligand binding, followed by autophosphorylation, followed by recruitment of signaling complex components (binding). Instead, curators should use a single activity unit with an MF term (in this case, GO:0004714, transmembrane receptor tyrosine protein kinase activity) which includes all of these mechanistic steps.

Exception: multifunctional proteins. In some cases, distinct catalytic activities are carried out by distinct domains of a protein, e.g. subsequent steps in a biochemical pathway, or histone reader and writer activity. In these cases distinct Activity Units for the same product should be used and causal relations be added to link them.

Connecting pathways

Each GO-CAM model should ideally correspond to a biological “pathway” or “module”. A module is roughly defined as a group of genes that function together in a subpathway, and that subpathway is reused in multiple larger pathways. Modules can be linked together into larger pathways via causal connections between reactions in different modules. Currently, the VPE does not allow causal connections to be made between distinct GO-CAM models. When a connection is desired, curators should include one additional activity (the one from the other model that is connected) in each of the two models, and make the causal connection in each of the two models. For example, upstream and/or downstream pathways can be connected by capturing a causal relation between the last activity of one pathway and the first activity of the next pathway. This allows the generation of global, whole cell-level metabolic maps.

{#heading}

Review status {#review-status}

Reviewed on 2025-05-20
Reviewed by Paul Thomas, Pascale Gaudet, Patrick Masson

Revision history {#revision-history}

Reviewed on
Reviewed by