Docker for reproducible computational research

Overview

Bioinformatics analysis usually involves a large number of software tools, reference data and pipelines used to elaborate the results. Reproducing the same analysis by other researchers is often a burden as many pieces of the puzzle are missing from the used methodology. While the raw datasets are generally available; a clear workflow/ pipeline detailing the results reproducibility is often missing. In order to achieve reproducibility in computational biology, publishing a clear commented source code is a crucial step, but this is not enough as in almost every case the working environments are not armed with the right tools and dependencies to run the code. The biggest obstacle in computational reproducibility would be to create a reliable, standalone, multiplatform and lightweight-working environment in which all the computational needs for a study are met. Virtualisation and containerisation are the two approaches to address this issue. While virtualization e.g. VirtualBox is an option, it is memory intense and computationally expensive with limited not scalable performance and usually difficult to couple with high performance computing platforms. Containerization e.g. Docker is a widely used as a lightweight fast and scalable alternative to Virtual machines as it communicates directly with the Kernel of the host operating system. It can easily be deployed on a high performance computing clusters or to a cloud based elastic computation center e.g. Amazon web services.

The Docker technology position itself as promising approach to computational biology research reproducibility by

• Saving time and expenses on human and computational resources allocated to already performed analysis
• Boosting communication between computational biologists working on similar topics
• Enhancing transparency within the community
• Granting open access computational knowledge to the community
• Building upon previous discoveries rather than building all over

Audience

This course is addressed to bioinformaticians and life scientists.

Learning objectives

During this one-day tutorial participants will practice basic Docker command line functionalities, eg setting up a Docker image, deploying images as “containers” and opening ports targeting pre-installed high throughput sequencing software tools. We will also introduce Amazon web services as cloud based tools hosting the pre-built Docker containers. The knowledge acquired by the participants in this tutorial should allow them to fetch and build reproducible workflows using Docker technology.

Prerequisites

Knowledge / competencies

Knowledge of the Next generation sequencing techniques is not required however Basic Unix command line knowledge is needed.

Technical

Participants should bring their own laptops with Docker installed and register to AWS.

Application

The registration fees are 60 CHF for academics and 300 CHF for for-profit companies. This includes course content material and coffee breaks.

Deadline for registration and free-of-charge cancellation is set is set to June 6 2018. Cancellation after this date will not be reimbursed. Please note that participation to SIB courses is subject to our general conditions.

You will be informed by email of your registration confirmation.

Venue and Time

Universität Bern, room 028 EG/West, Hochschulstrasse 4.

The course will start at 9:00 and end around 17:00. Precise information will be provided to the participants on due time.

Additional information

Coordination: Walid Gharib and Patricia Palagi

We will recommend 0.25 ECTS credits for this course (given a passed exam at the end of the course).

You are welcome to register to the SIB courses mailing list to be informed of all future courses and workshops, as well as all important deadlines using the form here.

For more information, please contact training@sib.swiss.