FND-CSC-Data models
Relational Data Models
(Relational data models - what, why, how)
Abstract:
Computational work with data often begins with data modeling: parsing facts about the world into a set of entities, their attributes, and their relationships. These are usually represented in a "relational data model". This unit introduces the concept, discusses pitfalls in creating such models and how they can be addressed, and practices designing and evaluating datamodels.
|
Objectives:
|
Outcomes:
|
Deliverables:
- Time management: Before you begin, estimate how long it will take you to complete this unit. Then, record in your course journal: the number of hours you estimated, the number of hours you worked on the unit, and the amount of time that passed between start and completion of this unit.
- Journal: Document your progress in your Course Journal. Some tasks may ask you to include specific items in your journal. Don't overlook these.
- Insights: If you find something particularly noteworthy about this unit, make a note in your insights! page.
Prerequisites:
This unit builds on material covered in the following prerequisite units:
Contents
Evaluation
Evaluation: NA
Contents
Task:
- Read the introductory notes on data models.
In the PDF notes, a protein data model is developed. You can access a sketch of the data model by clicking on the image:
Open the slide, save it, and edit it. (Or download to use any other drawing tool.) To practice data modeling, think about and try modeling the following extensions to the "proteinDB" model (in a new slide):
- Some of the proteins that you might want to store are transcription factors. A transcription factor has a canonical binding site sequence, and there are sequences it actually has been observed to bind to. The actual binding instances in specific locations may have genes associated with them, which encode proteins. You might come up with other facts that are important too.
- A "systems model" would group together a number of proteins to a system such as "G1/S checkpoint control", "cell-wall repair", "acid/base homeostasis" etc.: i.e. a set of proteins that collaborate towards a common goal. Within that system a protein performs one or more functions, that may be associated with specific states of the protein (like its intracellular location, or its post-translational modification state). Proteins may be structurally part of any number of systems, and they may shuttle between systems. Systems can overlap, and sometimes we might want to group systems in a hierarchical fashion.
- A protein-protein interaction database stores interaction information. Interactions may be observed by a number of different experimental methods and thus several different interactions may be reported for the same protein pair. Moreover, there may be meta-information, such as a confidence score that evaluates whether two proteins functionally interact with each other in the cell, rather than the observation being an experimental artefact. Some of the interactions may be between a protein and a complex, or between two complexes and can't be further resolved. But if a interaction is between two disinct proteins, and one of them is part of a complex, that too is importnat to know. Some interactions may be permanent, and some may be transient, i.e. depend on particular conditions.
- Sketch each of these datamodels on your own. Think about the principles that were discussed in the introduction. You will probably start by listing the entities first, then the attributes of the entities, then the relationships that you need to represent the facts. Don't forget the cardinalities.
About ...
Author:
- Boris Steipe <boris.steipe@utoronto.ca>
Created:
- 2017-08-05
Modified:
- 2020-09-20
Version:
- 1.1
Version history:
- 1.1 New schema sketch
- 1.0 First live version
- 0.1 First stub
This copyrighted material is licensed under a Creative Commons Attribution 4.0 International License. Follow the link to learn more.
