| Project | Rearrangements in genomes with unequal content |
|---|---|
| Date | 09/06/05 |
| Version | 1.0 |
| Purpose of Meeting | To discuss issues in prototyping |
| Supervisor present | Leong Hon Wai |
In the past week, I carried out a very simple experiment in which I stared with an ancestral genome (identiy permuation of order n) and randomly apply k reveral operations to this genome to get a new genome. This was repeated 3 times to obtain 3 genomes. Then for each of the blocks in the 3 genomes there is a probability p of it being deleted. This generates a set of 3 genomes with unequal content. Further processing on this data was reqired to produce a data set which can be used as input to the web-based MGR. The reversal distance of the ancestral genome returned by MGR was compared to the identity permutation and recorded for different settings of n, k and p.
The results confirmed what we would expect. As the p increases, the distance of the recovered ancestor gets further away from the identity permutation. Prof Leong suggests that we carry out more experiments to study this phenomenon.
| dist | ||
|---|---|---|
| p | k = 10 | k = 5 |
| 0 | 3 | 0 |
| 0.05 | 3 | 0 |
| 0.10 | 4 | 0 |
| 0.15 | 4 | 2 |
| 0.20 | 6 | 3 |
| 0.25 | 7 | 5 |
Prof Leong recommends that we make use of CVS when we are developing the implementation as this allows for effective version control.
As we need to follow the RAS methodology for the conduct of proper experiments, Prof Leong suggests we make use of an OO wrapper to provide access to the MGR code developed by Dr Bourque.
Prof Leong reminded us that we need to develop a plan before running any experiements. It is important to know what output should be produced by the experiment such that the data can be analysed and most importantly be reproduced later on. (refer to RAS methodology for running experiments)
Prof Leong suggested that we should prepare a proper minutes for every meeting, this will allow us to keep track of the important points that were raised in the meeting as well as allow Dr Bourque to follow the progress of the project.
Similarly Prof Leong recommends that we start to work on the project plan as well as prepare to give a talk about the general area of the project. The project plan will include information such as a proposed schedule, list of milestones, focus of the project, application of the results, list of project risks etc. He will sent a copy of the project plan to Dr Bourque so that he can check that the project is in line with his expectations.
I pointed out the the RNG in C/C++ is implementation dependent, this means that it might not be possible to reproduce the exact same experiment on a different machine even when using the same seed. Prof Leong recalled that a similar problem was brought up in the past and they have already developed a solution to the problem. Prof Leong will elaborate on how to overcome this issue in a future meeting.