T6: Inferring gene regulatory networks from high-throughput data with ISMARA

Organizers:

Mikhail Pachkov, Jeremie Breda, and Erik van Nimwegen (Biozentrum, University of Basel, and Swiss Institute of Bioinformatics)

Tutorial Summary:

Motivation 
By modeling genome-wide gene expression and chromatin state data in terms of computationally predicted binding sites, Motif Activity Response Analysis (MARA) allows automatic inference of the key regulators, their targets, and their interactions from high-throughput data in any system. In recent years MARA has been completely automated into an integrated system (ISMARA) webserver that allows any researcher to upload their data and obtain comprehensive predictions of key regulatory network structure in their data. The ISMARA system is quite sophisticated and provides users a large number of interactive possibilities to explore predictions and to generate new analyses of the data and even many experienced users are only aware of a fraction of the possibilities that the system provides. We here propose to provide an in-depth interactive tutorial of the ISMARA system.

Expected Goals
The attendants of the tutorial will learn how to use the system, what kind of data ISMARA is able to analyze, and obtain an in-depth exploration of all the analysis results that the system provides including:

  • What are the key regulators, their activities, and the expression profiles of these regulators?
  • What genes and pathways are targeted by each regulator?
  • What is the core network of interactions between key regulators.
  • What are the main regulators of a particular gene and how is this gene's expression affected by each regulator?
  • Exploring embedded links to the String and SwissRegulon databases.
  • How to average across replicate data and how to calculate contrasts between particular samples.
  • How to download comprehensive predictions and potential post-processing of these results.

At the end of the tutorial the attendants should have the expertise to perform sophisticated regulatory network predictions from RNA-seq or ChIP-seq data using the ISMARA system.

Level and intended audience
In principle the tutorial should be of interest to any researcher that aims to make computational inferences about gene regulation from gene expression and chromatin state data, including purely experimental researchers. It will be of particular interest to computational biologists and bioinformatic researchers that regulatory analyze gene expression data in their work. No specific bioinformatic skills are prerequisites although a basic understanding of gene expression analysis and transcription factor binding site predictions will be helpful. Attendents are expected to be familiar with the molecular biology of gene regulation in higher eukaryotes.

Prerequisites
Users should, whenever possible, bring their own laptop and data. Wireless connection will be needed for users to be able to interact with the system.

Tutorial Agenda:

Tuesday, September 12, 2017
Venue: Kollegienhaus, University of Basel

9:00 – 10:30 Introduction to Motif Activity Response Analysis. Transcription factor binding site predictions, processing of raw RNA-seq and ChIP-seq data, and the MARA model. Users that brought their own data can upload them to the webserver.
10:30 – 11:00 Coffee break
11:00 – 12:30 Using ISMARA: data types that are supported, and usage of the ISMARA client or direct upload. Overview of the basic results that ISMARA provides: Lists of key regulators, their activities, lists of target genes, and core networks.
12:30 – 13:30 Lunch break
13:30 – 15:00 Advanced interactive features of ISMARA. Exploring the regulation of particular target genes. Pathways targeting by particular regulators. Averaging across replicate samples. Calculating contrasts between subsets of samples.
15:00 – 15:30 Coffee break
15:30 – 17:00 Users explore results on their own datasets (or provided test datasets for users that did not bring their own data). Tutors provide guidance.

Tutorial Speakers:

Jeremie Breda is a PhD student in the PhD Program "Fellowships for Excellence". He works in Mihaela Zavolan's group and with Erik van Nimwegen. During his master in physics, he developed an enthusiastic interest in different application of statistical physic’s theory to biological systems. He first came to the Biozentrum to accomplish his master thesis working on the optimisation of the biophysical miRNA-mRNA interaction model MIRZA and then started his PhD in 2015. He's currently working, amongst other things, on adapting ISMARA for single-cell measurements.