BIO systems project

Bioinformatics Project: Defining a System

   

This course gives you a broad overview of bioinformatics principles, but you should also strive to explore one aspect of the field more deeply.

For your term project I would like you to identify a defined biological "system" - a set of genes that collaborate towards a shared purpose. We start by looking at biological processes, represented in the Gene Ontology (GO). From there we can find related processes, functions and cellular components. The problem and reason why we need human intuition to work out a systems definition based on this kind of information is that there are more aspects to a system than just the actual function: genes that are responsible for substrate import, biosynthesis of cofactors, signalling, regulation, constructing scaffolds etc. may also be part of the system. On the other hand some genes participate in a central role in making the process possible, but they provide this service to many other systems as well and are actually parts of a distinct but collaborating system. Membrane transporters might be an obvious example.

It is your task to manage this from the perspective of a biological expert and try to define inclusion/exclusion criteria as best as you can. While your "list of genes" is going to be interesting, compiling such lists can be automated. Thus the most valuable outcome of your project how you will address the task of defining your system boundaries.

In practice you should

• define the biological process you are interested in;
• collect all contributing genes as best you can, using a broad spectrum of literature comments and bioinformatics tools that we may have or have not covered in the course;
• develop unambiguous criteria for including or not including such genes in your system;
• provide an annotated list of included genes, and ones that you have excluded; and
• carefully document your efforts and results: the datasources, what procedures have been applied, how the results been accessed, validated and interpreted...

Ideally, your process would be defined at a level where the system that realizes it is comprised of some 20, 30 genes or so, not much more, to keep things manageable.

Open topic

The function you choose is open. I have posted a list of suggestions. However, you should ensure you don't choose the same process as someone else in the class.

First stage: Choosing a suitable process (5 marks max.)

To define a system, we will start from a biological process in the GO biological process ontology. I have excerpted a table of processes to get you started, explained the procedure in detail and worked it out in one example. You can find all of this here.

• Table of GO terms - choosing a process to define a system

Note that you are not constrained to start from a process in that table. If you are determined to work on a different human system, you are welcome.

The page also links to an example page on my Student Wiki. The example page illustrates what I expect from you for full marks for this stage.

More notes ...

Keep your systems manageable. When considering how many genes are associated with a system, check the taxon section of the relevant GO terms' statistic on QuickGO. The number of genes involved in the process in humans is likely as large as the largest number for ANY species - although many of the human genes may not have been annotated for that process (yet). For example, if the mouse (mus musculus) has 20 annotated genes and humans have only two, that probably does not mean humans can achieve with only two genes that for which the mouse needs twenty. Part of the next stage will be to attempt "annotation transfer" between orthologues. You will need to consider the genes individually...

Keep your systems simple. I would avoid choosing systems/processes that integrate sensory, nervous, hormonal and cellular components. This may become too complex. Narrowing it down, to a manageable "subsystem" is a valuable exercise in itself. Such a system may implement

• integrating input,
• transmitting input signals to their effectors,
• regulating the process,
• providing resources,
• defining setpoints,
• assembling or disassembling the system,
• mediating interactions with other systems,
• or similar...

(I'm just throwing these terms out there but I think we probably need to work out a systems roles ontology (SyRO) for the next stage, to have some context against which we evaluate the individual genes' roles.)

Spend some thought on naming your "system" well. For example a concept like immune response does not allude to why the system exists. I think naming the concept defense against pathogens captures this better.

We actually have an interesting situation. It is common for science to ask how questions, not why questions, because the why questions are thought usually not to have a scientific answer, i.e. they are not well posed in the sense that an answer might not exist, might not be unique, or might not be verifiable as being an answer. But we have discussed that evolution works by selecting from (neutral) variation according to an organism's fitness function. This allows us to formulate an answer to a why question: a system exists because it improves the organism's fitness function. In general we have no way of quantifying the fitness function - it represents a very high-dimensional multi-parameter optimization problem. But what we can observe is the existence of purifying selection. This gives us a rigorous, testable, scientific perspective: a system exists because it does something which results in traces of selection.

 

Second stage: Compiling a list of genes (12 marks max.)

The second stage of the project is for you to detail the roles that your system needs to work, and to associate genes with roles.

On one hand, you need to figure out how your system comes into existence, how it accepts substrates and/or information, how it transforms this input and how its output is generated. Consider that whatever is switched on, needs to be switched off again. And think clearly about the ultimate point of the system: why is it being selected for in the first place. The Systems Roles Ontology may help you, and if it does not match your needs for your system, contact me and we will improve the ontology.

On the other hand, you need to collect genes that contribute to those roles. All tools of bioinformatics are fair game for this: finding homologs, looking for information in PubMed, looking for similarity in GO, querying pathway databases, asessing protein-protein interactions etc. etc. You will probably amass a significant number of genes. But then it becomes important to draw the line: which genes are at the centre of your system, and which genes should really be part of something else. As you make these decisions and shape the boundaries, you should maintain an in and out list: genes that you keep in the system, genes that you declare as being outside and a note on why you made that decision. The latter is most important. At first, the goal is to describe the system, but the ultimate goal is to abstract the decision making process and automate it.

Just like defining how to tie a tie.

 

More notes ...

Review your system concepts. ...

Add genes ...

Compare against the role ontology. ...

Identify "adjacent" systems. ...

Define your system's boundaries. ...

 

Final stage: Documentation (9 marks max.)

The documentation must fulfill two aspects.

• First, your documentation must make your data and results reproducible. You need to specify the premises you started from and how you came up with them, and you need to specify the procedure through which you arrived at your conclusions. Put yourselves into the mind of a reviewer: are you providing enough information so that your (computational) steps can be reproduced? Are your source IDs specified? Your resources and programs? Have you made your R scripts available? The parameters for analysis?

• Second, your documentation must explain the rationale behind your procedure and conclusions. This is not so much what you did but why you did this, what was the logic behind a certain process or decision.

• Form is important:

• structure your project clearly, include a brief introduction and definitely include a meaningful conclusion;
• avoid jargon;
• make it easy to copy data for further analysis (no screenshots unless you are illustrating a Web-site or GUI);
• write complete sentences;
• do not plagiarize, but reference judiciously;
• make sure your references are complete and take advantage of the <ref> ... </ref> tags and the {{#pmid:1234567}} template.

Ask(!) if you are not sure about Wiki markup or formatting to achieve a particular layout.

Due dates

 

The function choice is due by the end of week 6.
The compilation of the list of genes and documentation are due before the Exam. If you need time beyond that data, you must notify me before the exam.

 

Late submissions

The time of submission is recorded with your edits on the Wiki and can be identified in the View history tab of a page: I will consider the last edit before the submission deadline for marking. However, if you want me to consider a later edit instead (i.e. "late submission" with the appropriate penalties), send me an eMail to that effect. If you don't email me, your mark from an incomplete submission will stand.

Please get your deliverables done early, I will be quite resistant to grant extensions for reasons that have to do with your normal, expected workload. If you want to, you can submit all phases of your project at any earlier date you choose - and get it done with. Be especially mindful of your other courses, and their midterm tests.

Just to clarify: "by the end of ..." means Tuesday at midnight. And yes, there will be penalties. Your final mark for the stage will be multiplied by the following factor for each day after the deadline on which it is submitted:

Marked on the ...

• first day after the deadline: marks times 0.9
• second day: 0.7
• third day: 0.4
• fourth day: 0.1
• fifth day and later: 0

 

Resources