Guide to Character Annotation
Contents
- 1 Curation Goal for June 15, 2009
- 2 PATO terms used for annotation
- 3 Character Annotation Examples
- 4 Refining terms by post-composition
- 5 Evidence Codes
- 6 How to write definitions for ontology terms
Curation Goal for June 15, 2009
The current goal for curation is to annotate all characters to a coarse level of entity and quality. For entity, we are picking entities (or short post-comps for processes, see below) and for quality we are using a small set of PATO terms (see table below). By curating all A-level papers to this level, we will focus on adding new entities to TAO, and categorize phenotype variation into broad categories which will guide future annotation development efforts.
PATO terms used for annotation
Curators are currently annotating systematic characters at a coarse level using a small set of quality terms from PATO. These terms represent higher-level concepts to describe phenotypic variation, and the Quality field in Phenex is currently limited to these terms.
| PATO term | Synonyms | Character description examples |
|---|---|---|
| shape | triangular, lobed, concave, interdigitated | |
| position | placement, location | horizontal, vertical |
| size | thin, large, decreased height | |
| structure | porous, non-porous | |
| composition | ligamentous | |
| ossified | ||
| cartilaginous | ||
| texture | smooth, wrinkled | |
| color | ||
| relational shape quality* | protruding into | |
| relational spatial quality* | anterior to, lateral to | |
| relational structural quality* | fused with, overlap with, separated from | |
| present | present in organism, presence | |
| absent | absent from organism, absence | |
| count | count in organism |
(*) Relational qualities describe phenotypes between two entities
Character Annotation Examples
The following are examples of character types commonly encountered in the systematic literature and how we annotate them using the EQ model in Phenex. Abbreviations: E, entity; Q, quality, RE, related entity, C, count.
1. Presence/absence
E: pectoral fin, Q: present in organism E: pectoral fin, Q: absent in organsim
If a character describes the presence or absence of a structure located on another entity, (for example, presence of teeth on a particular bone), first check to see whether the structure exists as a term in the TAO before creating a post-composition.
The statement "Teeth absent on basihyal bone" is annotated as:
E: basihyal tooth, Q: absent in organism
2. Well developed vs. small or absent entities
For example, a character involving the auditory foramen is coded with two states: 0, absent or small; 1, well developed.
State 0 is recorded as polymorphic in Phenex:
E: auditory foramen, Q: decreased size E: auditory foramen, Q: absent
State 1 is recorded as:
E: auditory foramen, Q: increased size
3. Use of monadic qualities
Monadic qualities are those that exist in a single entity, such as shape, and do not require another entity. For example, annotation of “sigmoid-shaped supraorbital bone” is entered as:
E: supraorbital, Q: shape
4. Use of relational qualities
Relational qualities are those that exist in an entity but require an additional entity in order to exist. No qualities that are children of monadic entities (e.g., size, shape) are relevant here. For example, annotation of “parietal fused with supraoccipital” is entered as:
E: parietal, Q: relational structural quality, RE: supraoccipital
5. Counts
Characters involving counts of entities are annotated using the “count in organism” quality. Values for counts are entered in the “count” field(*). Note that ranges and lower or upper bounds are recorded as follows:
E: vertebra, Q: count in organism, Count*: 33 E: vertebra, Q: count in organism, Comments*: 34-38 E: vertebra, Q: count in organism, Comments*: >38 E: vertebra, Q: count in organism, Comments*: =/>48
*Note: The Count field in Phenex is currently configured to only accept integer values. Record ranges or min/max values in the Comments field.
6. Annotation involving bone and cartilage terms
Authors sometimes refer to endochondral structures without specifying whether the structure is “bone” or “cartilage.” For example, the convention for a term like “epibranchial 2” or “basibranchial” is that it is composed of bone and when composed of cartilage an author will say “epibranchial 2 cartilage” or “basibranchial cartilage”. However, this is not universally followed and a curator must read the character description and examine associated text and figures to ascertain this. If after reviewing the publication it is still not clear, then the curator can use the “element” term for the structure. In general, the “element” terms should be used sparingly, and only when an author does not indicate whether a structure (e.g. “epibranchial”) is composed of cartilage or bone.
The following are a series of examples demonstrating the use of bone, cartilage, and element terms.
Example 1
Epibranchial 1: (0) present and ossified
E: Epibranchial 1 bone, Q: present
Epibranchial 1: (1) present and cartilaginous
E: Epibranchial 1 cartilage, Q: present
Epibranchial 1: (2) absent
E: Epibranchial 1 cartilage, Q: absent E: Epibranchial 1 bone, Q: absent
The curator should use both the cartilage and bone terms to annotate state 2 because the author clearly differentiates between the two.
Example 2
Epibranchial 1 (0) present, (0) absent. After careful reading of associated text in the publication, the curator cannot conclude that the author refers to bone or cartilage, and therefore uses "element" term:
E: epibranchial 1 element, Q: present or Q: absent
Example 3
Epibranchial 1 bone: (0) present; (1) absent
E: epibranchial 1 bone; Q: present or Q: absent
Example 4
Epibranchial 1: (0) triangular. After careful reading of associated text in the publication, the curator cannot conclude that the author refers to bone or cartilage, and therefore uses "element" term:
E: epibranchial 1 element, Q: shape
Example 5
Epibranchial number: (0) 3; (1) 4. After careful reading of associated text in the publication, the curator cannot conclude that the author refers to bone or cartilage, and therefore uses "element" term:
E: epibranchial element; Q: count; C: 3 or C: 4
Example 6
Epibranchial bone number: (0) 3; (1) 4
E: epibranchial bone; Q: count; C: 3 or C: 4
Refining terms by post-composition
**Note: for the current curation goal, we are creating very simple post-compositions for the presence or absence of processes/shelves/ridges/etc... on bones. For example, (process^part_of(entity)) is composed without reference to spatial orientation.
Oftentimes the need arises to use a more granular term than is already present in the ontology, whether it is the anatomical entity or the quality. For example, to prevent ontology "bloat", regions, margins, and projections of a bone are not in the anatomy ontology, but the bone itself is, and the concepts of margin, relative location of the margin (anterior, posterior, ventral, dorsal, etc), or bony projection are too, or are in other ontologies (spatial aspect, for example). Similarly, the directionality of a phenotype (such as a rotation, or curvature) isn't necessarily present in PATO, but the component terms necessary to express it are. The act of combining terms on-the-fly into cross-product terms is called post-composition, as opposed to pre-composed terms that are already in the ontology.
Post-composed terms can be created in Phenex following the genus-differentia principle of defining terms, where one term serves as the genus, which is then differentiated using a relationship and a differentia term. Unlike pre-composed terms, post-composed terms do not have an ID, and hence are "anonymous." Therefore if the same post-composition is used multiple times, it has the same semantics, but not the same identity. For example, if one wants to assign multiple annotations to the same process of the lateral ethmoid in the same specimen, using post-composed terms does not allow the identity of the anatomical structure between the annotations to be inferred.
The order in which the terms are composed is important if the composition relationship is not reflexive (a relationship is reflexive iff A rel B <=> B rel A). Most relationships are not reflexive. As a general rule, choose the more general part as the genus, and then use the relationship and differentia to narrow down. For example, for "process of the lateral ethmoid" use "process" as the genus, and use the relationship and differentia to narrow down which process of the many that are possible you mean, such as using part_of for the relationship and "lateral ethmoid" as the differentia (formally, the "process that is part_of the lateral ethmoid").
Note that semantically, the post-composed term is equivalent to a pre-composed term, provided the pre-composed term has both the inheritance relationship and the cross-product relationship properly recorded. For example, "process of lateral ethmoid" is-a "process", and "process of lateral ethmoid" part-of "lateral ethmoid".
Refining entity terms on-the-fly
To post-compose a term (for example, "maxillary process") using Phenex, first type in the genus term in the Entity field within the "Phenotypes" panel. Then open the Post-composition Editor box by right-clicking on the Entity cell (make sure that the entity cell is blue in color) :
Click on “Edit Post-composed Term.” The Editor box should now appear (see below). To post-compose the entity “supraorbital projection”, type “process” in the Genus field, click the “+” button to add a row in the table, type “part_of” in the relationship field, and type “supraorbital” in the differentia field. Click OK.
The post-composed term appears in the Entity field as:
process^part_of(supraorbital bone)
Refining entities on the fly using spatial terms
Often we will want to include spatial information in a post-composed term. The Spatial Ontology is used to post-compose terms related to bone margins, surfaces, or regions. For example, the entity “anterior process of the maxilla" is post-composed as follows:
Within the Entity field, type “process” and right-click the field to open the Post-Composition Editor box. Within this box, Click the + button to add relationship = "part_of" and differentia = "anterior region". Right click on the differentia field to open another Post-Composition Editor box to add second post-composed term for "anterior margin of maxilla" (see below). Click the OK button after you have filled out the genus and differentia for this nested composition.
Within Phenex, the post-composed term for process on the anterior margin of the maxilla appears as:
process^part_of(anterior region^part_of(maxilla))
As for semantics, identifiability, and rules for post-composing, the same applies as above for entity terms applies to spatial terms. In creating the nested term "anterior region of maxilla", start with the more general term (for example, the "anterior region") as the genus, then refine it using a relationship and a differentia term (for example, anterior region that is part_of the "supraorbital bone").
Evidence Codes
**Note: Does not apply to current curation goal
We record phenotype descriptions as properties of species, and annotations are assigned one of three evidence codes based on the level of evidence given by an author for phenotype observations. These specimen evidence codes are in an Evidence Codes Ontology that was developed by the broader biological community (see http://obofoundry.org/cgi-bin/detail.cgi?id=evidence_code). We have added evidence codes to this ontology, and we use the following in order below from strong to weak evidence.
Inferred from Voucher Specimen (IVS)
Used when an annotation is made on the basis of a phenotype description for a species or higher level group that is given by an author who explicitly references an observation of a voucher specimen(s). Voucher specimens are defined as those specimens with permanent museum catalog numbers. Thus it would be possible for a person to examine this particular specimen and observe the annotated phenotype.
- Note: if there is a matrix in the paper, the IVS evidence code is assigned to all annotations linked to the character list.
Traceable Author Statement (TAS)
The TAS evidence code covers author statements that are attributed to a cited source. Typically this type of information comes from review articles. Material from the introductions and discussion sections of non-review papers may also be suitable if another reference is cited as the source of experimental work or analysis. When annotating with this code the curator should use caution and be aware that authors often cite papers dealing with experiments that were performed in organisms different from the one being discussed in the paper at hand. Thus a problem with the TAS code is that it may turn out from following up the references in the paper that no experiments were performed on the gene in the organism actually being characterized in the primary paper. For this reason we recommend (when time and resources allow) that curators track down the cited paper and annotate directly from the experimental paper using the appropriate experimental evidence code. When this is not possible and it is necessary to annotate from reviews, the TAS code is the appropriate code to use for statements that are associated with a cited reference. Once an annotation has been made to a given term using an experimental evidence code, we recommend removing any annotations made to the same term using the TAS evidence code.
Nontraceable Author Statement (NAS)
The NAS evidence code should be used in all cases where the author makes a statement that a curator wants to capture but for which there are neither results presented nor a specific reference cited in the source used to make the annotation. The source of the information may be peer reviewed papers, textbooks, database records or vouchered specimens.
How to write definitions for ontology terms
Genus-differentia definitions
Term definitions in the teleost anatomy ontology (TAO) take the form of genus-differentia definitions: B is an A that has X. The term B is defined by its membership in higher category A and distinguished from its sibling terms by characteristic X. The following are examples of genus-differentia definitions in the TAO:
Antorbital: Dermal bone that is located on the anterior margin of the infraorbital series, dorsal to the first infraorbital and lateral to the nasal bone.
Dentary: Dermal bone that forms the anterolateral part of the lower jaw.
In example 1, the definition mentions the parent dermal bone of the term antorbital, followed by the characteristics that differentiate antorbital from all other dermal bones.
Post-composed terms (described below) are terms created on-the-fly at the time of annotation, and take the form of genus-differentia definitions.
