Help:Semantic Annotations

From Human Salivary Proteome Wiki

Semantic Annotations

In a typical wiki, information stored in text and files are easily accessible to human readers. Although it may be easy to retrieve individual facts, knowledge cannot be readily aggregated, especially from large datasets. Therefore, the Human Salivary Proteome Wiki places on top of the free-text structure a layer of "meta-data", known as semantic annotations, to make the information computer-interpretable. Creating these semantic annotations enables you to extract otherwise obscure knowledge from the data (even though there is some additional work to set up these annotations). The semantics features of the system come from an extension called Semantic MediaWiki (SMW).

Semantic annotations have three parts: a subject (whatever is being described), a value (something related to the subject), and a property (how the value is related to the subject). This is similar in many ways to a simple sentence in most languages. For example, consider the sentence "The short form of the Glucose-6-phosphate 1-dehydrogenase protein has a molecular mass of 59,256.3 Daltons." We could say that the subject is the a specific variant of Glucose-6-phosphate 1-dehydrogenase, represented by HSPW:PDB9827/1, the value is 59,256.3 Daltons, and the property is "Has molecular mass." The sequence is related to the number by the fact that the number is the sequence's mass.

In the wiki, every semantic annotation describes the page it is written on, so the subject is always taken care of: whatever page you put your annotation on becomes the subject. An annotation thus can be specified as follows:

[[property::value]]

In the Human Salivary Proteome Wiki, "Has molecular mass" is the property used for specifying the mass of a molecule. The value for this property must be a number, and the unit is assumed to be Daltons (Da). Therefore, from the example above, we could insert the annotation by adding the following to the HSPW:PDB9827/1 page:

[[Has molecular mass::59256.3]]

However, you usually do not need to worry about adding semantic annotations manually because the commonly used properties have already been embedded in most pages. When a page is created, the annotations are generated automatically.

You can read more about semantic properties and their types at the Semantic MediaWiki Web site.

Semantic Features on the Wiki

More than a million facts are stored in the wiki as semantic annotations. The wiki is in essence a large knowledge base of various information about the salivary proteome. To effectively disseminate the knowledge to a wide range of audience, the annotations are utilized in many ways:

Data tables

Many of the tables you see on the pages are generated dynamically by pulling data from various places (See Figure 1). This allows the data to always be up-to-date and the display can easily be modified based on the particular information need.

Fig. 1. A table from a Protein Cluster page created using semantic queries to display detailed information about members of the cluster.

Dashboards

To provide an overall picture of the information without drilling down to the details of individual pages, data are summarized and presented in charts and lists to answer common questions many users have (See Figure 2). Many of these dashboard figures can be found in the category page of the entities.

Fig. 2. A pie chart showing the different types of protein signatures detected in the saliva proteome.

Ad hoc queries

You can search for specific information by creating your own queries on the Special:Ask page. You can also access the page by clicking Search > Semantic search in the navigation menu. To learn more on how to construct a query, please see Help:Semantic Queries.

The ability to create your own semantic query is very powerful but is usually not necessary for most circumstances. The wiki has another advanced search tool that utilizes semantic annotations but is more user-friendly. Go to this interface by clicking Advanced search under Search in the navigation menu. For more information, please refer to Help:Searching#Advanced_Search_using_Semantics.

Browsing interfaces

One way to view all the semantic annotations on a page is through the Special:Browse interface (Figure 3). The interface can be accessed by clicking the 'Browse Properties' link under the More tab on top of the page. The browser lists properties and their values defined on the page. In addition, it also has the ability to display incoming annotations, where the entity of interest is the value of an annotation from another page. Links to other pages have eye-shaped icons next to them. Clicking on these icons will take you to the browse interface for the linked page.

Fig. 3. The browse interface displays both incoming and outgoing properties of a page.