Action: Writing about my perspective of writing the abstract and data summary and having gone through the first module of 20.109.
Motivation for me taking the class:
At first, I was really excited. All the work that had been done over the past few months from DNA extraction to gel electrophoresis to transforming competent cells would be culminated in a roughly 10 page report on our research. Although I have been working on a biological engineering UROP over the last summer, I was really impressed by how 20.109 labs worked: there was always big EMPHASIS on understanding the protocol and each step that went into task of understanding the microbiome and how our research was related to that. For example, I never really took the time before to understand chaotropic salts and their role in DNA extraction.
What did I want to write about:
For the longest time, I had convinced myself that our research pertained to the relationship between the seagull gut microbiome composition and host susceptibility to avian influenza virus. As a researcher, we would be doing the dirty work of dealing with avian cloacal samples that could somehow be related to avian influenza virus (although we were told that samples that contained human pathogenic influenza were excluded from the study). While the cause was worthy, the limitations placed on our research seemed to be due to our dearth of lab experience or because our class was only an introduction lab class or maybe because we needed to be specially trained to deal with infectious viruses... I sort of missed the purpose of the experiments.
What I ended up writing about:
There was a little disappointment when I realized that our topic covered what I felt to be a more simple scope of inferring characteristics of the host from the microbiome. However, I realized that all research starts at a foundation and builds up from there. My conclusions from our research focused more on the correlation between the microbiome composition and gender and geographic location of the host, but future implications of modulated microbiome therapies and microbiotic markers for disease susceptibility cause me to realize that there is a lot of potential in the field and study of microbiomes and phylogenetic analysis. How I got to these conclusions is another story.
Troubles I ran into with data analysis:
There was a lot of data collected from just 8 seagulls, an entire 109 16S sequences that were all roughly 1400 base pairs in length. While the hard work of filtering out the 16S DNA amplicon and running BLAST analysis was split among the class. The task of using and interpreting MEGA and UniFrac data and even individual parsing of the data proved rather tedious and led me to the realization that scientific research is difficult. Especially telling a story from the resulting data felt like being told to solve a 100-piece puzzle except instead of all the pieces being unique jigsaw pieces, all the pieces were identical 1''*1'' squares and you're told that some of the pieces are missing so you need to infer what the puzzle is a picture of without all the pieces. So I went to work: observing the strange trees and trying to make sense of what patterns and information could be parsed from each graph and whether the heat maps told any story at all.
I stared at my computer screen feeling utterly perplexed as to why none of the data seemed to be making sense. I saw some sort of relationship in the cluster sampling between male gulls from Worcester and female gulls from Carson Beach, but all of the other points seemed to be outliers or seemed to serve the only purpose of destroying any theories I had with regards to microbiome composition and host gender and geographic location. Having then exhausted what I could from the UniFrac website, I began to explore the 20.109 website to see if I had missed something that could explain my lack of tools in data parsing. Suddenly in a "Huzzah!" moment, I located the Spring 2015 tab-delimited text file containing all the results of the BLAST queries on closest matching strain to the 16S sequences and their corresponding gull number. If I could systematically organize the data, maybe I could have a better idea of the composition of the gull microbiomes being compared and make better inferences of the relationship that microbiota had with their host. Now how do you change a text-delimited file back into a excel file and how can I most efficiently sort and organize all these 109 results and how best could I represent this data in a figure and...
Finishing at the point of daybreak:
After an exhaustive process of chugging through and gaining better understanding of the microbiome, I looked at my report. After several weeks of weekly coming into labs and getting nervous that something could go wrong, (which it did, huge thanks to the teaching staff and assistants for being so awesome with troubleshooting) here was a paper that represented that work in the form of a schematic, an image of a 1400 base-pair band of DNA in a gel after gel electrophoresis along with other trees and figures that tried to tell the story of a class of MIT sophomores in course 20 who tackled the question of what a microbiome was and how to even quantitatively/qualitatively describe and interpret variation of microbiome structures as seen across coastal and inland communities of seagulls. There was a feeling of pride and satisfaction as I looked through my entire work. In a sense, I took the puzzle pieces, found more pieces and then started reasoning out a possible picture and started building towards that design and adjusting as needed until the final piece was crafted. Possible pieces may have been put in upside down or at the wrong angle, but that is hopefully where constructive criticism can come in to put pieces in more optimal places.
Expectations of the future: