Homework 4

Designing ontologies using Protege

Due: Thursday September 12

Turn in written answers to questions 1-6 in class and email your revised ontology to bright at cs.pdx.edu before class on September 12.

The purpose of this assignment is to become familiar with ontology management using Protege and gain some experience viewing and modifying an existing ontology.

Part 1. Getting started

Download and install Protege on your preferred platform at:

http://protege.stanford.edu/download/release/full/

IMPORTANT: Make sure you choose either "Basic+OWL" or "Everything" for your installation, since we will be using the OWL plugin for this assignment.

For those who prefer to work in the lab, an installation is available on the Linux machines (fab01-fab08) in the Intel Lab in FAB 55-17. Type ./stash/bright/Protege_3.1/Protege and Protege should run. Email bright at cs.pdx.edu if there are any problems.

Part 2. Using Protege

Do the Protege tutorial at:
http://protege.stanford.edu/doc/tutorial/get_started/get-started.html

Note that this tutorial is for Protege-Frames and the rest of the assignment uses Protege-OWL, but it will help you become familiar with the interface.
A (slightly out of date) Protege Owl Tutorial is available at
http://www.co-ode.org/resources/tutorials/ProtegeOWLTutorial.pdf
If you haven't already, read chapter 3 of this document (also available at http://www.cs.pdx.edu/~howe/cs410/owl_overview.pdf). This is only 4 pages, and provides a brief introduction to OWL.
You only need to look at pages 16-33 of the tutorial for this assignment, which are available at:

http://www.cs.pdx.edu/~howe/cs410/owl_tutorial.pdf
Do this part of this tutorial to familarize yourself with Protege-OWL. Note that for instruction 2 of Section 4.3 on page 21, instead of the Wizards menu you should select "Tools-\>Quick Owl-\>Create Multiple Subclasses".

Part 3. Viewing and modifying an ontology

Download the amino acid ontology available at:
http://www.co-ode.org/ontologies/amino-acid/2005/10/11/amino-acid.owl
To open this file in Protege, select "New Project...". A dialog displaying a list of file types will appear. Select the "OWL files (owl or rdf)" option and check the "Create from Existing Sources" box at the top, then click the "Next" button. On the next dialog, click on the grey box in the upper right corner to browse for files, then select the file containing the amino acid ontology. Click on the "Finish" button and a new project containing the amino acid ontology should open.
Recall from the biology portion of the course that there are 20 different amino acids, and each is denoted by a different letter of the alphabet.

A classification of the 20 amino acids is available at:

http://prowl.rockefeller.edu/aainfo/chemprop.gif

Look at the graph labeled "physio-chemical" on the left. Each amino acid is in one or more circles, depending on which class or classes it belongs to. A circle that is completely contained by another circle denotes a class that is a subclass of another (for example, the tiny amino acids are a subset of the small amino acids). Two circles that are completely disjoint denote two disjoint subclasses (for example, an acid might be either aromatic or aliphatic, but not both).
Question 1 (5 points):
Click on the triangle next to AminoAcid to show its subclasses and click on AliphaticAminoAcid. What are the properties of an aliphatic amino acid?
Question 2 (5 points):
Look at the list of subclasses of AminoAcid. Do any of these subclasses look out of place compared to the others? Why?

Question 3 (5 points):
Suppose you wanted to know which of the 20 amino acids are aliphatic. How would you find out by viewing this ontology in Protege?

Question 4 (5 points):
Which classes that appear in the classification diagram in http://prowl.rockefeller.edu/aainfo/chemprop.gif are not listed as amino acid subclasses in the ontology?

Right click on the subclass AromaticAmino acid and select "Add subclass". Select the H amino acid. Next right, click on the PositiveChargedAminoAcid, select "Add subclass", and select H again.

Question 5 (5 points): Click on H and look at its asserted conditions in the window on the left. What are the inherited conditions? (listed at the bottom, below necessary conditions).

Question 6 (5 points): Click on each of the two hasCharge Positive rows under asserted conditions and delete these two rows. Do the inherited conditions below change? Why or why not?

Ontology Modification (20 points):

Make at least two changes to the amino acid ontology to make it reflect the classification in the figure at http://prowl.rockefeller.edu/aainfo/chemprop.gif.
Possible changes you can make include:

-Adding amino acid subclasses that are not currently included as subclasses of amino acid. Make sure you add the appropriate property values as well.

-Explicitly adding a subclass relationship between two classes.

-Explicitly making two classes disjoint.

Write a few sentences describing the changes you made and briefly justify your changes. Save your ontology (include your name in the file name) and email the files your_name.pprj and your_name.owl to bright at cs.pdx.edu by the deadline.