|
Weeks:
1
2
3
4
5
6
7
8
9
10
Week 6
Jun 26 - Jul 2
On Friday our team met to discuss the project. Bob had some exciting news. We would actually be submitting a paper based on this project for the PSB conference this year. But there was a problem - the deadline for submission was less than three weeks ahead and we had absolutely no result. We went over what each of us had to do. Jian would be working on the web crawler for gathering information on ADRs from social networks. Laura would annotate some of the ADRs in Jian's collected data. Bob and Ryan would build a system for automatically extracting all mention of adverse event in Jian's data using Laura's annotations. I would try to find some meaningful way of analyzing the drug/gene/ADR relationships. Obtaining results for my part was especially important since the conference was on personalized medicine. After the meeting I explained to Bob what I've been doing till now. Bob looked over my results and agreed that the current technique didn't look too promising. He suggested we try to figure out a new approach.
Over the weekend I met Gazi's roommate- Nabil who was from Bangladesh too and the three of us basically hung out. Professor Gonzalez had also e-mailed me a paper on comparing networks and had given me a book on managing databases and I read through them. The paper on networks was interesting but I wasn't sure how to apply it to our project. Managing databases was a good read. I've been mainly doing manual work all along and I was completely sick of it. I realized that a lot of things I've been doing will be far easier if I simply used MySQL.
Over the week I wrote the introduction for the paper we would be submitting. I also continued explaining to Bob my part of the project and we tried to come up with some other plan. I was very frustrated by then. I had spent hours on the project and it seemed to be producing no result. I also missed coding a lot. All I've been doing was work with excel. If our hypothesis had worked we could have moved into something more exciting but the work I was left doing was tedious and boring. I wondered whether I had made a wrong decision by choosing a new project. It would have been so much easier if I had just gone ahead with one that people were already working on or had a better idea about. Something new in the field of text mining may also have been easier since most of the people in the Diego lab worked with text mining. Most importantly I would have been writing programs like CS students do instead of spending hours copy-pasting in excel!
Bob was very supportive all throughout. One of the things that kept me going was a saying of Thomas Edison which Bob quoted for me: "I have not failed. I've just found 10,000 ways that won't work." Bob also explained that research is not about coding. It is about doing whatever you need to do to make your project work and if that involves tedious, manual labor then you have to do that as well.
By the end of the week we finally decided to follow up on a technique that seemed to have worked okay. For say 10 ADRs, I would find all the drugs suspected to cause it. For each of these drugs I'd find the gene targets. For each of the genes I'd find the pathways they are associated with. In this way I'll be able to find an ADR-pathway link. I can then rank the pathways by counting the number of times this pathway was associated with the ADR and finally check whether drugs that affect these pathways cause similar ADRs. For doing all these I'll be using MySQL.
|