Workshops & tutorials

Keywords: algorithmic fairness, bias, gender dimension, edi

Description

In this workshop, we would like to investigate the hidden biases that might exist in our bioinformatics daily practices, be it in the use of particular data repositories, the curation process, the way we design experiments and implement models, or the analysis approaches we undertake and the interpretation of their results. We will focus on sex and gender biases, since their consideration for studies in human and other species is nowadays highly encouraged, for example in grant applications (see the EU MSCA guidelines).  

Davide Cirillo, head of the Machine Learning for Biomedical Research Unit at the Life Sciences Department of the Barcelona Supercomputing Center (BSC), will be giving a 20min introductory talk on the topic. Davide’s research focuses on computational methods for precision medicine with a special emphasis on machine learning, network science, and ethics of artificial intelligence.

In small moderated subgroups, and drawing from experience brought by SIB members, we will then discuss questions such as "are protein and gene annotations unbiased"? "Is personalized medicine accounting for sex and gender biases by design"? "How can we test machine learning models for fairness"? 

By the end of the workshop, we should have identified potential hidden gender biases in bioinformatics research and services, and come up with ideas on how to start addressing them. The outcome of the discussions will lay the ground for future actions at SIB (e.g., specific trainings, written guidelines, …) helping bioinformaticians to integrate the gender dimension early on in their research programs, ultimately improving the scientific quality and the societal relevance of our work.

Audience and requirements

Anyone working with data will be valuable to the discussions, e.g. those generating the data, those analysing and modeling it and those annotating/curating it.

Organizers

SIB Diversity working group: 

• Aitana Lebrand (Chair)
• Maïa Berman
• Ute Roehrig
• Julien Roux
• Gerardo Tauriello
• Robert Waterhouse

Maximum number of participants: 25

Description

Sequencing-based epigenomics data are of enormous value to the scientific community. They contain the answers to tens of thousands of questions researchers all over the world are asking every day, potentially making a large number of new experiments obsolete. However, as long as epigenomics data are difficult to find and use, carrying out a new experiment may remain a faster path to the answer to many bench biologists. This workshop focuses on the encyclopedic value of epigenomics data, not on the holistic, systems-level perspective. Obstacles to findability are mainly related to the so-called "metadata crisis", which is synonymous to absence or poor quality of sample annotation. The solution to this problem consists of metadata standards and investment in data curation. Usability is hampered by the fact that most data are not available in an immediately usable, processed form. Downloading and processing large raw data files is time-consuming and computationally expensive. For look-up use, data visualization is key. In many cases, a quick look at a genome browser window provides the answer to the user's question. Therefore, an existing technical solution to this problem would be to provide all data in a browser viewable track format. 

Specific aims of this workshop are:

• Learning from participants about common practices and challenges in re-using published epigenomics data
• Discussing FAIRness issues in epigenomics with colleagues
• Raising awareness of FAIR principles among younger scientists
• Initiating a SIB focus group to continue the discussions initiated at the workshop
• Information on related initiatives: SwissBioData ecosystem, FAIRtracks

Audience and requirements

Everybody with an interest in or working with epigenomics data. PIs, postdocs, PhD students, data curators, core facility members.

This is not a workshop for participants who would like to learn more about a research field they are not familiar with. However, a mixed crowd is desirable. Participants should have working experience in at least one area relevant to the workshop (experimental epigenomics, data analysis and visualization, biodata curation, gene regulation ontologies) or be active in promoting FAIR principles.

Participants should bring a device to access the internet, preferably a laptop.

Organizers

• Philipp Bucher
• Michael Stadler

Maximum number of participants: 50

Keywords: Genome sequencing, genomic epidemiology, emerging viruses, phylodynamics, phylogenetics

Description

The COVID-19 pandemic demonstrated the importance of genomic epidemiology and it now forms an integral part of public health responses. To be useful for public health interventions genetic characterisation should follow within days of identifying probable cases, however standard genomic surveillance in Switzerland has a turnaround time of 1-2 weeks and there is no established strategy for the rapid, targeted sequencing and analysis of newly emerged viruses or virus variants during the first weeks of an outbreak.

In this workshop, we will discuss the challenges of implementing a rapid genomic surveillance strategy that can be activated in case of emergencies. Our discussions will focus on three target areas: rapid data generation, practical analyses and effective communication. After a brief introduction, delegates will divide into focus groups based on interests and expertise. Within each group, discussions will concentrate on characterising the problem, current strategies, state-of-the-art and areas for improvement. At the end, groups will reconvene and each group will have a chance to present their conclusions.

We hope to use these discussions as a jumping off point to develop a policy document with a set of actionable recommendations. All participants will be invited to contribute to this document.     

Audience and requirements

Researchers at all career stages are welcome. We are hoping for a mix of graduate students, early career researchers and established professors. It would be helpful if participants are familiar with any or all of genomic epidemiology, public health, sequencing technologies or infectious disease dynamics.

Participants should bring a device to access the internet, preferably a laptop or tablet.

Organizers

• Louis du Plessis
• Chaoran Chen
• Adrian Lison
• Cecilia Valenzuela Agüi

Maximum number of participants: 50

Description

This workshop offers an overview of good practice in experimental design for high-throughput experiments. Particularly we focus on pitfalls in the design of projects (e.g. too low sample size or sequencing coverage, or sample batches overlap with biological groups), how to successfully plan an experiment and how to deal with challenges, e.g. batch effects or confounding factors. General concepts and pitfalls will be illustrated and discussed with participants, with showcase examples from the field of compound screening.

Audience and requirements

Basic theoretical knowledge in biology/biochemistry/computational analysis. No particular computational or wet lab experience is necessary.

Organizers

• Michael Prummer
• Eva Riegler 
• Franziska Singer

Maximum number of participants: 40

Keywords: Epigenetics, high-throughput sequencing, gene regulatory networks, transcription regulatory factors, chromatin dynamics

Description

Do you have epigenomic data that measures changes in genome-wide chromatin state across a set of conditions and want to know which regulators and regulatory interactions are driving these changes? Our laboratory has developed CREMA: Cis-Regulatory Element Motif Activities, a tool that models epigenomic data (i.e. ChIP-seq, DNase-seq, or ATAC-seq) in terms of computationally predicted binding sites of transcription factors, to infer the key regulatory interactions that are driving changes in chromatin state. CREMA is implemented as an online tool that takes raw sequencing data as input, performs all the modelling of these data in a completely automated manner, and provides comprehensive predictions of the regulatory networks and regulatory interactions through an interactive online interface. In this tutorial, we will introduce users to the CREMA theory behind the analysis, and explore in detail how it can be used in practice. CREMA has a large number of interactive possibilities to explore predictions and to generate new analyses of the data which will be demonstrated in the tutorial. The main objective of this tutorial is to allow attendees to develop an in-depth understanding of the variety of analyses that this tool provides, and how to optimally use it for answering specific questions about the regulatory network operating in the user’s system of interest.

Audience and requirements

The tutorial is targeted to a broad audience. This tutorial should be of interest to all people keen on inference of gene regulation from chromatin data including computational biologists, bioinformaticians and experimental biologists. Since no specific bioinformatic skills are required in order to perform the analysis, purely experimental researchers with no data analysis background are also encouraged to attend.

The participants are expected to be familiar with the molecular biology of gene regulation. No specific bioinformatics or data analysis skills are required.

In order to optimally benefit from the tutorial, users are strongly recommended to bring their own laptop. In addition, although users can explore the results on provided example datasets, we highly recommend that users upload their own data a few days in advance of the course, so that they may explore the results on their own data. In our experience this strongly improves the impact of the tutorial.  

Organizers

• Erik van Nimwegen
• Mikhail Pachkov

Maximum number of participants: 25

Keywords: workflow management system, Snakemake, reproducibility

Description

With the current state of research in biology, virtually all biologists have to use or implement workflows to analyze data and produce figures, but most were never taught how to do that properly. Consequently, it's not uncommon for analyses included in published articles to be extremely difficult - sometimes impossible - to reproduce. This lack of reproducibility has become a pressing issue, and strict reproducibility guidelines are now being implemented by more and more journals and reviewers to tackle the problem. In this context, we would like to offer an informal introduction to Snakemake, a workflow management language / system, and Conda, a package manager, which can be used together to implement reproducible and portable workflows. The workshop is aimed at beginners; we will focus on explaining all the basic concepts required to implement simple workflows, and we will briefly introduce a few advanced topics that are particularly useful for common use cases.

Audience and requirements

Beginners who would like an introduction to Snakemake. No prior knowledge required, but a little experience with command line / Conda / Python will help.

Users will need to bring their own computer with Conda and Snakemake installed and set up properly (instructions will be provided ahead of time).

Organizers

• Romain Feron
• Antonin Thiebaut

Maximum number of participants: 20

Keywords: protein structure prediction, quality estimation, AlphaFold, SWISS-MODEL

Description

During the last 2 years, the field of protein structure prediction has been revolutionized by end-to-end prediction techniques based on Deep Learning. In this tutorial, we will introduce participants to the most recent advances in protein structure prediction. We will start with a short presentation introducing the current state of the art. A hands-on tutorial will then focus on web based tools that can go from one or more input protein sequences to an accurate 3D model of a single protein chain or a protein complex. Topics covered will include homology modelling with SWISS-MODEL and deep learning powered modelling with AlphaFold. At the end of the tutorial, participants should be able to execute their own protein modelling tasks and judge the quality of the output of the available methods. We expect the schedule to consist of ~30min for the introductory presentation, ~10min to introduce the tutorial and the largest amount of time for hands-on exercises.

Audience and requirements

Any life science researcher interested in using protein structures in their work. No prior knowledge is required beyond basic knowledge of proteins.

Users are strongly recommended to bring their own Laptop. Currently a Google account is needed to access some web-based modelling options (based on Google Colab). This may change until June 2022.

Organizers

• Gabriel Studer
• Joana Pereira
• Gerardo Tauriello

Maximum number of participants: 25

Keywords: single-cell transcriptomic, gene expression analysis, coarse-graining

Description

Single-cell RNA sequencing (scRNA-seq) technologies offer unique opportunities for exploring heterogeneous cell populations. However, in-depth single-cell transcriptomic characterization of real biological tissues, like tumors, often requires profiling tens to hundreds of thousands of cells. Such large numbers of cells represent an important hurdle for downstream analyses, interpretation and visualization. In this tutorial, we will present different approaches to simplify single-cell RNA-Seq data by merging highly similar, and possibly biologically redundant cells into metacells. These will include the SuperCell method developed in our lab. We will demonstrate that metacells not only preserve but often improve the results of downstream analyses including visualization, clustering, differential expression, cell type annotation, gene correlation, imputation, RNA velocity and data integration. By capitalizing on the redundancy inherent to scRNA-seq data, metacells significantly facilitate and accelerate the construction and interpretation of single-cell atlases, and allow users to analyze RNA-Seq profiles of millions of single cells in a few hours on a standard desktop.

Audience and requirements

The tutorial is targeted to bioinformaticians analyzing single-cell RNA-Seq data. Participants are expected to be familiar with basic programming in R and have a basic understanding of RNA-Seq gene expression data.

Users are strongly recommended to bring their own Laptop (R studio. with enough memory - ideally at least 8GB).

Organizers

• David Gfeller
• Mariia Bilous
• Aurélie Gabriel
• Léonard Hérault

Maximum number of participants: 30

Keywords: science communication, outreach, SNSF Agora grant

Description

Are you passionate about science and want to share it with others? Do you think more scientists should be speaking up about societal issues that can be impacted by their science? Do you wish you could explain what you do in an easy and fun way? If you’ve answered yes to any of these questions, this workshop is for you. We will present different aspects of our ongoing Agora grant on evolution (including an activity with marbles and memes) - see https://lightofevolution.org and https://ohmygenes.org. By the end of the workshop, the participants are expected to: 

- describe the work they do and its impact on society
- activity that introduces the general public to their work
- prepare a brief communication around the activity

The workshop will end with a general discussion on participating in science outreach activities or writing an SNF Agora grant. 

Audience and requirements

Everyone working at the SIB. No prior knowledge requirements.  Users should, whenever possible, bring their own laptop.

Organizers

• Marie-Claude Blatter
• Natasha Glover
• Monique Zahn

Maximum number of participants: 50