This was a very exciting week! I am continuing to work on the project that I started last week in Matlab. On Monday, I completed doing a correlation of the data and began working on finding the significant edges of the network. I found the p-values of the edges in order to determine which ones are significant, however, the p-values that were found along with the amount of significant edges indicated that the significant edges could all be noise. Therefore, two more steps had to be taken in order to verify which edges are truly significant and not just noise. First, I completed a false discovery test, which is a more stringent way of obtaining significant edges. However, the false discovery rate did not work (which sometimes happens) so the next step was to determine if there are any clusters in the data. If there are clusters, the edges are most likely significant and not noise.
On Monday, I also got to experience my very first lab meeting! It was very exciting! One of the post-docs in the lab, Sarah talked about her research and we then spent an hour talking about an article. A week before the lab meeting, Sarah chose the article because it sounded interesting. However, in the lab meeting, Sarah along with the other lab members tore the article apart! Sarah explained how the research conducted in the article sounds interesting but the methodology is not correct. I thought it was so interesting how critical they were of the article because for me, any published research sounds good! I definitely learned to look at everything with a critical eye. On Tuesday I was not at the lab.
On Wednesday, I continued finding the components from the data (the clusters). I then did a Permutation test. In order to do the Permutation test, I had to randomize 1000 matrices of data and in order to do this, I wrote my very first loop! A loop repeats an order of steps and arguments a defined amount of times (a for loop does that to be exact).
Dr. Bassett is running, for the first time, a Networks Visualization Program along with Sara Hodgson from Incompra Design this summer for 6 weeks. In the program they selected art majors from local Universities to use their artistic skills to create networks, while also learning more about biological systems and complex networks. From Thursday, June 26 to Wednesday, July 2, the mornings will include lectures from various professors and post-docs from around Penn who are working with networks. Dr. Bassett invited me to come to the lectures!
On Thursday morning, I listened to two lectures given by Dr. Bassett about first, networks in general and second, about neural networks and her lab in particular. We finished at around lunch time and I went back to the lab where my lab was having a U.S.A. vs. Germany lab party with food and a big projector. After the party, I continued working on the permutation test and determining the significance levels.
On Friday morning, I listened to two more lectures. One from Ann Hermundstad, a Physics post-doc at a lab that uses networks in order to analyze the interactions between neurons in the retina. The lecture was incredibly interesting. The second lecture was from Matthew O'Donnell a researcher at the Annenberg School of Communication who is part of the Communication Neuroscience Lab. The lab focuses on how the brain of an individual can predict the actions of a population. They also examine how people's social network can determine how influenced they are by a group's opinion. That lab showed me a different aspect of networks science. On Friday afternoon, I finished collecting the significance level p-values for the different data sets.
Networks science is one of the most interesting aspects of science and it is amazing to think that I barely new it existed in the beginning of the school year. I am very excited to see what the next 7 weeks have in store for me!
No comments:
Post a Comment