At last week’s meeting, we dug into the computational portion of the project, discussing different ways we could run our algorithm on the gene lists and how that would affect our results. We also discussed a few different methods for running on our gene lists – one method involves a support vector machine and the other is an iterative method of computing scores, shown below.
Biology, 2011.
Alex was able to track down some software for the former method from the autism paper that inspired our project. One of our tasks was to try to get this software to work on our current gene lists – unfortunately, we hit a bit of a wall with this method.
For the second method, we were able to successfully run it on Alex’s gene list. I also had the task of running the iterative method on a small toy example, which was a small graph consisting of 10 nodes, 11 edges, 1 positive, and 1 negative. Anna informed me that when she did it, the scores converged after about 27 time steps, which matches the results that I got. We will go over the results in this week’s meeting and discuss what our goals for this week are.