Edited by Katyanne M. Shoemaker
Would you be able to explain your research to an audience of academics from all different disciplines, in just 3 minutes, with only one slide? That is the premise of a competition called the Three Minute Thesis (3MT). 3MT was created at the University of Queensland, Australia in 2008, and it has been performed at the University of Massachusetts Dartmouth, USA since 2011. (Details about the competition can be found here: http://www.threeminutethesis.org).
While thinking about the goals of this blog, I decided to participate in the competition this year, as it does exactly what we try to do here: talk about science to a diverse audience while keeping it interesting and educational. I signed up thinking only about the training; I would have to prepare and memorize my text and then deliver the presentation (in English!). Of course, I also had the ultimate goal of winning (who turns down a chance to earn $1,000?).
Unfortunately, I did not get rich on April 29th, 2015, but as expected, it was great practice and lots of fun. It was interesting to watch presentations about the research from various fields: engineering, arts, administration, etc. There were nervous people and people who seemed to have come straight from a theater stage. You can watch some videos of previous years by visiting the following website: http://www.umassd.edu/graduate/spotlights/three-minutethesiscompetition/.
You can read the transcript of my talk and learn more about my research below:
Many people do not know, but fisheries management is not just based on adult population data. It is also important to study early life stages for better stock management. For example, as fish eggs are usually spawned in the water column, knowing when and where they are helps to define spawning sites and periods.
But, before doing any kind of fish studies, it is necessary to know who they are. Fish egg identification is time consuming and difficult. After sampling on board, you need to sort all of the fish eggs from the plankton sample, using a microscope. Sorting the eggs from the family I am studying is easy because their eggs have an ellipsoid shape.
The problem is reaching the species identification. As each group presents different size and shape, the identification has previously been done by manually measuring each egg and then counting.
In my doctoral thesis, I want to verify long-term fluctuations in the abundance and distribution of eggs from a fish named Argentine anchovy on the Brazilian coast. This small fish is one of the most common fisheries resources in Argentina and Uruguay. At the Brazilian coast they haven’t been commercially fished yet, but some studies have suggested that Argentine anchovy can be sustainably fished in Brazilian waters.
Coming back to my thesis, when I mentioned that I am studying long-term fluctuations, I didn’t mention that by long-term I meant 40 years of data, totaling almost 2000 samples. That is a huge amount of samples and it would take my whole PhD period just to identify all the eggs. The solution was to create a faster and more accurate methodology, so I did it.
I used a digital camera attached to a microscope to image the eggs, and using the photos, I got the measurements. After that, I created a model that automatically gave me the counts of eggs within each species. This new model has over 90% accuracy and can be used by any researcher to optimize their time and effort. In the end, besides taking four years to identify the eggs for my thesis, I identified more than 100,000 anchovy eggs in just one year, allowing enough time to continue my research project.
If you are interested in this methodology, the paper is already in publication, and it can be accessed in the following link or requested by email.
See you soon.