The most notable change of the second semester is that we are going to study in a new school Stockholm University, and to learn programing in python, because we will use it in all of our four courses, namely Introduction to Bioinformatics, Project in Molecular Life Sciences, Biophysical Chemistry and Comparative Genomics.
(The content of Introduction to Bioinformatics and that of Biophysical Chemistry is well covered in my blogs Course Review: Introduction to Bioinformatics and Course Review: Biophysical Chemistry, therefore I will rather not spend any time here)
Project in Molecular Life Sciences
As its name suggested, it is a 100% project based, lecture-free exam-free course, which implies that all grades (from A to F) will be given to the performance, including:
- A full written report
- A final presentation of the project
- Participation and attendance in weekly seminar
- Quality of codes
As there are no lectures, the whole class were assigned two teaching assistants who can offer guidance throughout the 1.5 months of the course.
Then, what are we going to do in these 1.5 months? The task that we were given is to construct a predictor that returns the secondary structural elements of a protein sequence. How many possibilities are they? Those elements can be alpha-helix/beta-sheet/coil, membrane embedded/solvent-exposed, extracellular/intracellular……That’s why besides looking into the biochemical properties of amino acids, we have to employ the powerful machine learning tool.
It is my first time to take a project-based course on my own. Suddenly, I am “liberated”: no more lecture or exam, what I need to do is designing my predictor and coding for it. But you can imagine that it is never easy:
Firstly, I have to familiarize myself with the machine learning tool sk.learn that I use, which is a time-consuming process for any neophytes. After successfully running sk.learn for the first time, I have no time to celebrate, because the accuracy is not satisfactory, I have to dive down into my hundreds of codes, look for possible explanation, find out a solution, and do the accuracy test again ——- when the algorithm gets more complicated, the running time for accuracy test accelerates: it takes a night for checking only 1/6 of sequences that I have! It is an endless circle until the last minute before submission.
Sisyphus, is me
On the other hand, I feel that I am so lucky to be in my current class: we are adversaries, but at the same time friends and comrades. We gave comments, advice and criticism on each others’ work, also share information, for example when we run the accuracy test in order to save time. When I look back into the six weeks I spent in the computer lab, I start to realize that how robustly we were growing in this phase. We will grow even stronger.
In addition, my classmate Carolina also has a lot to say on this course on her Karolinska blog, here is the link: https://studentblogski.wordpress.com/2017/03/30/how-to-survive-2-months-of-programming/