Genome-wide association studies (GWAS) are still the major work horse for identifying novel genes that are associated with disease risk or certain traits. In this talk I will cover the basics of genomics (genome organization, genetic variation, etc.) and introduce how genetic variation is measured at low cost today (sanger sequencing, genotyping chips, next generation sequencing, …). Next, I will introduce the underlying principles of genome-wide association studies, basic tests, genetic model (assumptions), power analysis and the rationale of ever growing studies. I will introduce the motivation behind ‘imaging genetics’, i.e., investigating the genetic basis of phenotypes that we can drive from images. Further, I will cover the practical aspects of actually conducting GWAS, this will entail basic steps like how the data actually looks like (popular file forms), cleaning data, quality control, imputation of missing SNPs (and why this is possible), confounding factors such as population structure (and how to address this), and checking your GWAS results for ‘correctness’.

This talk aims at covering the statistical background required to perform association analysis in typical imaging-genetics studies. We will introduce the notion of statistical association, and highlight the standard analysis paradigm in univariate modeling. We will then explore multivariate association models, generalizing to high-dimensional data the notion of statistical association. In particular, we will focus on standard paradigms such as Canonical Correlation Analysis (CCA), Partial Least Squares (PLS), and Reduced Rank Regression (RRR). We will finally introduce more advanced analysis frameworks, such as Bayesian and deep association methods. Within this context we will present the Multi-Channel Variational Autoencoder, recently developed by our group.

Understanding the mechanisms underlying brain structural and functional variation is essential for advancing neuroscience. Magnetic resonance imaging (MRI) can be used to derive metrics of brain structure and function and offer a powerful method to assess disease burden in the brain.
Genetic drivers of brain differences are important to identify as potential risk factors for heritable brain diseases, and targets for their treatment. Imaging genetic studies have found that, as with other complex traits, a single common variant explains less than 1% of the population variance, despite accounting for a large fraction of the variance in aggregate. Therefore, successful studies require tens of thousands of scans, as well as an independent sample for replication. Large-scale consortia in the field of neuroimaging genetics, including the Enhancing Neuro Imaging Genetics through Meta Analysis (ENIGMA) consortium, have identified common genetic variants that have small but significant associations with variations in brain structural morphometry. Large-scale biobanks have been amassed tens of thousands of MRI scans of individuals from a single scanner for genomic discoveries, yet, replicating effects and ensuring generalizability of findings to current scanned populations require assurance that the brain measures being studied are reliably extracted across a variety of possible MRI scanning paradigms.
We will show the most common techniques and tools used by the neuroimagers community to extract anatomical and functional features from MRI data. Using examples, we will provide instructions for quality control and statistical methods to assess the robustness of the extracted features that will be used as phenotype for the genetic studies.

Clustering is a crucial procedure in exploratory data analysis. Given some data, clustering, combined with some visualization, is probably where start to fish for any useful data patterns. I will carry out a hands-on workshop where we will dive into some of the most famous clustering approaches. These will include hierarchical clustering, k-means, DBSCAN, and network-based clustering. We will also combine clustering algorithms with dimensionality reduction and embedding approaches, and learn about principal component analysis, multidimensional scaling, and t-SNE. We will learn how to apply these techniques to images that we will profile with deep learning models. During the workshop, we will use Orange, a data mining framework, and participants are welcome to download and install it from http://orange.biolab.si to follow along.

This course aims to facilitate the use of computing demanding applications in the field of NGS data analysis. The main feature to perform a correct experimental design will be explained. Then, the tools for RNA-seq data analysis will be detailed considering the following steps: quality control, normalisation and data reformatting, selection differentially regulated genes/microRNA, multiple testing and biological interpretation. All these steps will be explained from the theoretical point of view followed by a set of computational exercises to analysis a set of RNASeq data.
The analysis will be performed based on Docker4seq package. This package uses docker containers that embed demanding computing tasks (e.g. short reads mapping) into isolated containers. This approach provides multiple advantages:

• user does not need to install all the software on its local server;
• results generated by different containers can be organized in pipelines;
• reproducible research is guarantee by the possibility of sharing the docker images used for the analysis.

In the last years, the study of brain functional connectivity (FC) has become an increasingly active field of research providing novel and crucial insights in resting-state as well as during tasks. FC can be derived from different non-invasive magnetic resonance imaging (MRI) modalities, such as fMRI based on the BOLD contrast or Arterial Spin Labeling (ASL). Classical connectivity analyses allow detecting regions with similar behaviours or coherent signals, indicating their membership to the same functional network. In particular, FC can be computed in both time and frequency domains exploring different approaches (e.g., seed-based correlation, independent component analysis, spectral coherence). More recently, new approaches based on graph theory have been proposed for extracting significant aspects of the network and quantitatively characterising its global organization. In this lecture, I will present an overview of the main non-invasive functional MRI modalities and their applications for task and resting-state paradigms. Moreover, I will illustrate the associated methods of analysis for both brain localization and connectivity estimation, giving a glimpse of their main advantages/disadvantages.