Back

Bioinformatics

The Experience of Computer Cluster

Computer Cluster
Photo: Changsheng Li

Changsheng Li, student of the second course of BDBI, explains how a computer cluster works. It is a very useful tool that makes the work easier (sometimes just possible), not only for students but for bioinformaticians in professional research and production practice.

In the 2nd trimester of the second year of the Bachelor’s Degree in Bioinformatics, when the world is still struggling with the pandemic, we were introduced to the use of a computer cluster in the course Structural Bioinformatics. There are many calculations in Bioinformatics that cannot be performed in laptops but require specialized machines with large memory, large disk space and many processors. In Structural Bioinformatics we perform molecular dynamics simulations of proteins. These simulations offer a computational microscope permitting to visualize the motion of molecular systems at high spatial and temporal resolutions, enabling the study of dynamics of biological systems at atomic resolution in a way that is not accessible experimentally.

Via computer cluster, we are able to submit task queues on the cloud and run them multi-threadedly and continuously over hours, if not days. Furthermore, we don’t need to do any maintenance. And we can change to more advanced equipment easily. And when all the work and studying have to be migrated online and accomplished remotely, cluster makes it possible to maintain the academic standard of practical sessions.

The ESCI-UPF is currently providing us, as students of the Bachelor’s Degree in Bioinformatics, a cloud base cluster under Amazon AWS. This service has made us possible to run tasks like state-of-the-art protein molecular dynamics simulations. Such tasks require a high level of computing power and continuity over days and would never be accomplished by a traditional PC. But now we can work on a more familiar device. It somehow alleviates the complexity of work and yet offers a more powerful computing power. Especially, given the current situation, suitable for coping with potential COVID-19 contagion.

It was not until the introduction of the use of a computer cluster into our course that brought the practicability of certain courses like Structural Bioinformatics up to a brand-new level. Certainly, we could learn all this knowledge purely theoretically, so I was surprised to get in touch with a computer cluster with a bachelor’s degree. It makes certain over-loaded practical tasks and training practical and meaningful.

Computer cluster

Photo: Puzzle Factory

Other than Structural Bioinformatics, personally, whenever programming in C++, as a user of Mac OS, I’d encounter problems. The situation would worsen when fulfilling a half-finished file from professors instead of writing one from scratch. Because instead of a standard compiler G++, Mac OS only offers a simulative alternative: Clang. It doesn’t function perfectly especially when optimizing the running time of a certain program. Sometimes the program would even run illogically slower after proper optimization. I decided to make full usage of the cluster and run my programs on it. It worked perfectly and I also learned a lot during such interaction with the cluster.

Given the experience above, I could see that when it comes to cross-platform consistency in teamwork, a computer cluster also happens to be a great solution. And not only just in this degree, but it also plays an important role in professional research and production practice. As a future bioinformatician, I really appreciate such an experience. And for the new students in the years to come, I’d encourage you to take it seriously and maximize its potential, as it gives us more competitiveness in the future.

We also recommend you