Course: Analysis of single-cell RNA-seq data

Single cell RNA-sequencing (scRNA-seq) is providing an unprecedented granularity in the study of tissues and diseases. Once we abandon the assumption of a sample composed of average cells, with an average gene expression, the full heterogeneity of the different cells composing a tissue can emerge. Isolating specific expression profiles, usually through unsupervised clustering, it is therefore possible to characterize different cell populations or states (for example, separating stromal from immune cells). Once isolated, these populations can be further clustered in a nested fashion, to let sub-population emerge (for example, characterizing M1 versus M2 macrophages). In turn, these cell populations offer a tremendous potential to characterize prognostic or diagnostic biomarkers, unveiling molecular aberrations by comparing tissue under different conditions. Cell composition, i.e., the relative proportion of different cell types, constitutes a first example of these biomarkers. Another example of biomarkers are cell states, i.e., altered regulatory programs with identified cell types as the cells adapt to transient environmental changes, such as senescent phenotypes. A third biomarker example comes from gene expression, as scRNA-seq allows the detection of signals from rare cells. In contrast, a faint signal would be impossible to detect via bulk tissue RNA-seq, as it would get diluted by averaging it over the whole sample expression. Furthermore, it will be possible to decouple a gene signal in the sample down to the different contributions of diverse cell populations.
We expect the biomarkers from scRNA-seq analysis to be adopted in the future clinical practice, especially to characterize profiles of precision or personalized medicine. Despite the temptations offered by of-the-shelf, one-solution-fits-all scRNA-seq approaches and pipelines, extracting and interpreting this kind of data comes with some caveats and peculiarities. In this workshop, we will learn how to conduct a scRNA-seq data analysis, and we will address pitfalls and problems with real-data examples.

Learning objectives for the course
The attendee will learn the basic principles of single-cell RNA sequencing technologies, and how to conduct a typical analysis of single cell RNA-sequencing data. She will learn to perform quality control; to handle outlier; to perform data enrichment; to isolate and label cell populations; to find statistically significant biomarkers; and to communicate results by visualizing informative and convincing figures.

Class format
Each class is four hours in length. The course is designed as a lecture-laboratory, with guided in-class exercises to illustrate topics covered in the lecture, as they are covered.

20 December: 14:00 – 18:00

Introduction: Techniques for single cell RNA-sequencing
• Seurat single cell analysis pipeline: Pros and Cons
• Data extraction and format
• Data projection and enrichment
• Clustering and Biomarkers
– Break –
Hands on exercises

21 December: 14:00 – 18:00

• Dealing with multiple sets
• Dealing with batch and lane effects
• Outlier detection
• The importance of visual data representation
• Case study: cell stemness
– Break –
Hands on exercises

Course participation allows for 1 Credit for the students of the PhD Program in Health Technologies, Bioengineering and Bioinformatics.

Room D8, Department of Electrical Computer and Biomedical Engineering, University of Pavia
via Ferrata, 5; Pavia