taxon rule for tissue development

[gocentral]

Hi All,

Should the terms under Tissue development and its children get a taxon rule? http://amigo.geneontology.org/amigo/term/GO:0009888 Thanks, Rama

Reported by: rbalakri

Original Ticket: geneontology/ontology-requests/10848

[gocentral]

• assigned_to: David Hill

Original comment by: ukemi

[gocentral]

Hi Rama. What rule would you use? Since tissues are found in both plants and animals, we can't make an only_in rule.

Original comment by: ukemi

[gocentral]

Hi,

Minimally it shouldn't be used to annotate yeast gps.

R

Original comment by: rbalakri

[gocentral]

• status: open --> closed-works-for-me

Original comment by: ukemi

[gocentral]

Added never in Saccharomycotina

Original comment by: ukemi

[gocentral]

Would it be accurate to say unicellulars don't have tissues ? For eg bacteria and archaea should also be excluded, I would say. Probably all fungi as well, and Dicty.

Pascale

Original comment by: pgaudet

[gocentral]

Hi All,

The fungi you may be most familiar with are never described as having a tissue. Fungal tissues do exist, especially in fliamentous fungi which generate a variety of differentiated cell-types that are included in reproductive tissues. Filementous fungi, while considered a uni-cellular organism is capable of producing tissue, especially related to sexual differentiation.

http://www.britannica.com/EBchecked/topic/279327/hymenium Written by The Editors of Encyclopædia Britannica

hymenium, a spore-bearing layer of tissue in fungi (kingdom Fungi) found in the phyla Ascomycota and Basidiomycota. It is formed by end cells of hyphae—the filaments of the vegetative body (thallus)—which terminate elongation and differentiate into reproductive cells. The hymenium may also contain support cells known as cystidia.

Another fungal tissue reference: http://simple.wikipedia.org/wiki/Fungus Symbiosis

Symbiosis means living together. Lichens are a symbiosis between a fungus and an alga or bacterium. In this partnership the algal cells live inside the fungus tissue. The end result is a new mat-like life-form which clings to rock and other surfaces. About 20% of all fungi are lichenized.

There are many more peer-reviewed publications that refer to fungal tissue. I don;t have time to create an exhaustive list when you can do the google search too.

Never in Saccharomycotina should be much more carefully reviewed for accuracy before implementation, please. Perhaps it's best to have never in S. cerevisiae until a full understanding of fungal tissues has been acheived.

Diane

Original comment by: dinglis

[gocentral]

From wikipedia: http://en.wikipedia.org/wiki/Saccharomycotina

Saccharomycotina is a subphylum of the phylum Ascomycota (fungi which form their sexual spores in sac-like asci), and consists of yeasts - they form no ascocarps (fruiting bodies), their asci are naked, and they can reproduce asexually by budding.[2]

The only class in this subphylum is Saccharomycetes.[3]

It includes the well-known Baker's Yeast Saccharomyces cerevisiae[3] and the genus Candida which infects humans.

So the taxon restriction works for Saccharomycotina but not for a lot of other fungi.

Original comment by: dinglis

[gocentral]

In re-reading what I just quoted, the accurate statement is Saccharomycotina can make tissues, the subphylum Saccharomycetes of which S. cerevisiae is a member does not.

So the restriction of the term in Saccharomycetes is what would be accurate.

Original comment by: dinglis

[gocentral]

I did not make the restriction for all fungi. I did make the restriction for taxa that I thought consisted only of single-celled organisms. I did add it for archaea and bacteria. I did not add it for dicty based on PMID:22930590. Here are the restrictions so far:

relationship: never_in_taxon NCBITaxon:147554 ! Schizosaccharomycetes relationship: never_in_taxon NCBITaxon:2 ! Bacteria relationship: never_in_taxon NCBITaxon:2157 ! Archaea relationship: never_in_taxon NCBITaxon:33630 ! Alveolata relationship: never_in_taxon NCBITaxon:33682 ! Euglenozoa relationship: never_in_taxon NCBITaxon:38254 ! Glaucocystophyceae relationship: never_in_taxon NCBITaxon:4891 ! Saccharomycetes relationship: never_in_taxon NCBITaxon:554915 ! Amoebozoa

Original comment by: ukemi

[gocentral]

Note that I moved the Saccharomycotina one down to Saccharomycetes based on Diane's suggestion.

Original comment by: ukemi

[gocentral]

Replace

relationship: never_in_taxon NCBITaxon:2 ! Bacteria
relationship: never_in_taxon NCBITaxon:2157 ! Archaea

With

relationship: only_in_taxon NCBITaxon:2759 ! Eukaryota

Better to have at least one positive constraint

"Tissue", like "organ" is a pretty wishy washy term. It's not clear if we're specifying true biological constraints here on constraints on how terms are used by different communities.

Within the metazoa, it's common to say that porifera (sponges) and trichoplax lack "true tissues" - e.g. https://en.wikipedia.org/wiki/Trichoplax - e.g. trichoplax has epithelial cells in layers but the layers lack a basal lamina which would be a defining feature of a true epithelial tissue. It's not clear that other communities use tissue in this restricted sense. If tissue is defined to be any clump of cells then all animals have tissues, as do many fungi, maybe algal and microbial mats too.

The CARO def of 'portion of tissue' is what we should use for GO:

Anatomical structure, that consists of similar cells and intercellular matrix, aggregated according to genetically determined spatial relationships

I'm not sure how to interpret this in the different fungal cases.

In any case I think we should just place a "do not annotate directly" tag on wishy-washy upper level terms and place proper constraints on more meaningful subterms, derived from uberon (which has lots of taxon constraints for metazoa) and PO and FAO (which need to have TCs added).

Original comment by: cmungall

[gocentral]

On further thought, I think we need a stronger mechanism here

If you look in AmiGO you can see the IBA annotations propagated from this Opisthokont node: http://www.pantree.org/node/annotationNode.jsp?id=PTN000215025

The non-metazoan annotations seem unjustified to me. Either the common ancestor used these genes for tissue development (unlikely) or there was convergence (intriguing but requires more evidence).

If we believe a process arose multiple times (which is necessarily the case for a wishy-washy definition of 'tissue development') then we should record this and use this for paint inference. E.g. if a process is homologous within taxon node T, then propagations within T or subnodes is valid, but not outside T. Similarly if a process evolved multiple times within T then propagation at the level of T would require additional justification.

Bgee have these kinds of evolutionary annotations on Uberon classes, could also be extended for processes

See https://sourceforge.net/p/bgee/wiki/Similarity%20annotation/

Original comment by: cmungall

[gocentral]

Hi,

In response to Diane's comment : I am not sure the meaning of tissues in fungi is the same as that in plants, and that's also different in animals. We need to be careful not to merge those concepts together just because the same terms are used.

-Pascale

Original comment by: pgaudet

[gocentral]

is_a: GO:0048856 ! anatomical structure development relationship: never_in_taxon NCBITaxon:147554 ! Schizosaccharomycetes -relationship: never_in_taxon NCBITaxon:2 ! Bacteria -relationship: never_in_taxon NCBITaxon:2157 ! Archaea relationship: never_in_taxon NCBITaxon:33630 ! Alveolata relationship: never_in_taxon NCBITaxon:33682 ! Euglenozoa relationship: never_in_taxon NCBITaxon:38254 ! Glaucocystophyceae relationship: never_in_taxon NCBITaxon:4891 ! Saccharomycetes relationship: never_in_taxon NCBITaxon:554915 ! Amoebozoa +relationship: only_in_taxon NCBITaxon:2759 ! Eukaryota

I must admit, I am a bit concerned about adding these kinds of constraints for terms like tissue development because at some point they may limit how people can use the GO to perform evolutionary studies. If we get too conservative, we could miss interesting results.

Original comment by: ukemi

[gocentral]

How so? If someone comes up with experimental evidence (or strong human-reviewed phylogenetic evidence) that contradicts any of these we don't throw the experiments out, we change the constraints. But it's almost moot here as 'tissue development' is too vague to have any interesting results.

Original comment by: cmungall