Add text to Canto explaining that GO annotations capture the 'evolved function of gene product'

[CuzickA]

After discussion with @ValWood today we thought that this need to be made clearer in the tool whilst making GO annotations.

Val to add further detail to ticket thanks.

[ValWood]

Here is a draft:

It deals with the most common areas of confusion:

1. Effectors are not GO activities (because they represent a range of activities)
2. Ee don't annotating the effector to the outcome in the host if the host detects it (annotating evolved function)
3. Explain why effectors, as single step, are represented as processes.

molecular function

A molecular function (MF) is a catalytic (e.g. protein kinase activity, chitinase activity) or binding activity, or any other activity that occurs at the molecular level. Molecular functions usually correspond to single steps performed by gene products.

"Effector" is not included in GO as a molecular function, because effectors have a large range of activities (e.g. chitinase inhibitor, protease inhibitor, or most frequently "unknown"). If the effector MF is unknown annotate only to a biological process term "GO:0140418 effector-mediated modulation of host process by symbiont" or one of it's more specific descendants.

Start typing a molecular function in the search box (type at least 2 characters). If you do not find the term you are looking for with your initial search, begin with a broad term (~transporter, transferase activity~ enzyme inhibitor,ANOTHER?). More specific terms will be suggested allowing you to refine your search iteratively before making your final selection ~(examples of more specific terms are: adenylate cyclase activity or biotin transporter activity).~ Gene products may be annotated with multiple molecular function terms.

Biological process

A biological process is series of events accomplished by one or more ordered assemblies of molecular functions, such as cell cycle regulation, ion transport, or signal transduction. ~It can be difficult to distinguish between a biological process and a molecular function, but the general rule is that a process must have more than one distinct step.~ Effector-mediated processes often 'single-step' but they are classified as a process to group the entirety of 'effector-mediated' biology as a single GO term, and to allow the expression of the regulated host processes.

Start typing a biological process in the search box (type at least 2 characters). If you do not find the term you are looking for with your initial search, begin with a broad term (effector-mediated... ADD A COMMON NON EFFECTOR EXAMPLE) More specific terms will be suggested allowing you to refine your search iteratively before making your final selection. Gene products may be annotated with multiple biological process terms.

GO captures the evolved roles of proteins. Therefore effector recognition and activation of the host immune system by a host resistance gene should not be annotated as the role of an effector (unless annotating a neurotrophic effector)- always consider the 'evolved role' of the pathogen.

[jseager7]

@ValWood Should these changes be applied to all versions of Canto, or just PHI-Canto?

Note that the text describing the annotation type is configurable for each Canto instance – in this example, that is the paragraphs beginning with "A molecular function is a catalytic … or binding activity" and "GO molecular function terms represent activities".

However, the latter paragraphs that describe how to select the terms (and give broad and specific term suggestions) are set in Canto's code, so the only thing that can be customised there are the term suggestions, and the annotation type name (all of these are derived from the configuration file). If you change any other text in the latter two paragraphs, the change will apply to all versions of Canto – unless we add another configurable 'slot' that can hold more custom text.

[ValWood]

This is only for PHI Canto

[ValWood]

Could you make the example options configurable in the boiler plate part? I think then we could deal with the rest in the instance specific part with some rejigging.

Note that this is currently a draft, but it would be useful to enable the changes now.

[jseager7]

This is only for PHI Canto

Oh, I missed the label on the issue. Whoops.

Could you make the example options configurable in the boiler plate part?

Do you mean the broad and specific term suggestions? They're already configurable. Here's what PHI-Canto's configuration currently uses for molecular function annotations:

broad_term_suggestions: "transporter, transferase activity"
specific_term_examples: "adenylate cyclase activity or biotin transporter activity"

[ValWood]

OK. I think we can probably rearrange to fit the current configuration options.

[jseager7]

@CuzickA does this still need doing? If the text above is still applicable then I can try to fit the text into the structure required by Canto.

[CuzickA]

@jseager7 I think the above text would still be useful to add to PHI-Canto.

[CuzickA]

@jseager7, I think this text would still be useful to add to PHI-Canto.

[jseager7]

@CuzickA Thanks. I've re-read the issue and it looks like I still need some suggestions for more appropriate term suggestions for the GO Molecular Function annotation type in PHI-Canto.

For Molecular Function, Val suggested 'enzyme inhibitor' but needs another term.

For Biological Process, Val suggested 'effector-mediated modulation of host process by symbiont' but also needs a common non-effector example.

[CuzickA]

begin with a broad term (~transporter, transferase activity~ enzyme inhibitor,ANOTHER?).

How about 'kinase activity' for the MF broad term search?

begin with a broad term (effector-mediated... ADD A COMMON NON EFFECTOR EXAMPLE)

How about 'biosynthetic process' for the BP broad term search?

@ValWood, what do you think?

[ValWood]

Sounds OK, I might be inclined to go with "catalytic activity" rather than kinase activity since most people will find "kinase" and catalytic activity is broader.