Welcome to the Reproducibility Tutorial and Challenge!

Open science and reproducible research are common goals across research communities. Openness and reproducibility facilitates scientific advancement and has the potential to accelerate discoveries. However, reproducibility is a complex issue -- even within the constraints of computational biology. New programs are developed and old ones are updated or abandoned. Computers change and improve, as does the technology used to produce sequence data. There is a recent push for transparency in research, with the sharing of raw data and of “in-house” scripts used for analysis increasingly encouraged or required by scientific journals. During this course you will be exposed to a survey of bioinformatic approaches to deal with a variety of data types. For your final project, you will attempt to reproduce the bioinformatic analyses and findings from a recent publication.

Note:

Participation in the challenge is entirely voluntary, but doing so could help the community better appreciate the state of reproducibility in the field of bioinformatics!

Introduction: BVCN Reproducibility Challenge

Goals of the Project

• Dive deeper into the world of bioinformatics through understanding established research
• Practice structuring a full computational project
• Get a sense of the level of detail in computational biology that helps (or hinders) reproducibility
• Learn pointers for how to make your own projects open source and reproducible

Challenge Organization

Because we recognize that during this time, many of you may be swamped with commitments and personal concerns, while still wanting to get hands-on experience with bioinformatics skills, we have designed the BVCN Reproducibility Challenge to have something for everyone. While we encourage those with the time and motivation to embark on the full Reproducibility Challenge detailed on this site, we have also designed a walk-through tutorial example of a mini reproducibility challenge using an established paper, which can be found in the first half of the site.

Reproducibility Tutorial

The reproducibility tutorial is a low-commitment way to see reproducibility in action. The first half of the website guides you through the process of reproducing a bioinformatics project, in this case a transcriptomic study of cold acclimation in larval zebrafish (Long et al., 2013).

The site is designed so that at the top of each page, you should see a button that says "Interact with Binder", as well as a download symbol to the left.

Clicking the "Interact with Binder" button will allow you to work with the code all online, without having to download anything. You can also work with the code by downloading it off of the website using the download button, or (which is easier), by cloning the Github repository.

Reproducibility Challenge

While the reproducibility challenge is designed to work for people who have a little more time to invest, you should have everything you need to complete the challenge from the tutorial and the instruction set under the subheading "Reproducibility Challenge" in the left sidebar. Participating in the challenge could truly push the science of bioinformatics forward, by providing a sense of how many papers are reproducibible and what the stumbling blocks are. The more participants we have, the wider a range of bioinformatics problems we can address!

Developers

This challenge was originally developed by Harriet Alexander, Carolyn Teppolt, and Maria Pachiadaki for a course in the MIT-WHOI Joint Program. This challenge for the BVCN is being developed, moderated, and maintained by Harriet Alexander and Arianna Krinos, with help from Ben Tully, Maria Pachiadaki, and Mike Lee.

1. Long, Y., Song, G., Yan, J., He, X., Li, Q., & Cui, Z. (2013). Transcriptomic characterization of cold acclimation in larval zebrafish. BMC Genomics, 14(1), 612.