Fri
20
Mar
Seminars and Conferences
Statistical inference of transcriptomics data
On Friday, 20 March 2026, at 3 p.m., will take place a seminar entitled Statistical inference of transcriptomics data organised by the Department of Mathematical Sciences "Giuseppe Luigi Lagrange"-DISMA, with professor Simone Tiberi, University of Bologna.
Abstract
The course will introduce basic biological concepts, and three types of transcriptomics data: bulk, single-cell and spatial. First, it’ll describe pre-processing steps (alignment, quantification and normalization) and exploratory plots (correlation plot, heatmap, PCA/MDS/UMAPs). Then, it’ll focus on statistical models for analysing transcriptomics data (notably, the negative binomial and the Dirichlet multinomial), and describe various types of analyses: differential gene expression, differential alternative splicing, eQTL and sQTL, pathway analyses, clustering of cells (single-cell data) and spots (spatial data), and identification of spatially variable genes. Finally, it’ll show how to visually compare distinct models' results via ROC curves, TPR vs. FDR plots, and the Venn diagram. We will also analyse real transcriptomics data via the R software language, and apply some of the approaches seen during the course.
Speaker: Simone Tiberi, University of Bologna
Biography
Simone Tiberi is an Associate Professor of Statistics at the Department of Statistical Sciences at the University of Bologna. He is an applied statistician, interested in mathematical modelling of biological data, typically employing Bayesian hierarchical methods and latent variable approaches. He primarily develops statistical methods for analyzing omics data, with a particular focus on isoform-level inference and alternative splicing processes; his methods are then distributed open-source as Bioconductor R packages.
The seminar is open to everyone and will take place in person.
Abstract
The course will introduce basic biological concepts, and three types of transcriptomics data: bulk, single-cell and spatial. First, it’ll describe pre-processing steps (alignment, quantification and normalization) and exploratory plots (correlation plot, heatmap, PCA/MDS/UMAPs). Then, it’ll focus on statistical models for analysing transcriptomics data (notably, the negative binomial and the Dirichlet multinomial), and describe various types of analyses: differential gene expression, differential alternative splicing, eQTL and sQTL, pathway analyses, clustering of cells (single-cell data) and spots (spatial data), and identification of spatially variable genes. Finally, it’ll show how to visually compare distinct models' results via ROC curves, TPR vs. FDR plots, and the Venn diagram. We will also analyse real transcriptomics data via the R software language, and apply some of the approaches seen during the course.
Speaker: Simone Tiberi, University of Bologna
Biography
Simone Tiberi is an Associate Professor of Statistics at the Department of Statistical Sciences at the University of Bologna. He is an applied statistician, interested in mathematical modelling of biological data, typically employing Bayesian hierarchical methods and latent variable approaches. He primarily develops statistical methods for analyzing omics data, with a particular focus on isoform-level inference and alternative splicing processes; his methods are then distributed open-source as Bioconductor R packages.
The seminar is open to everyone and will take place in person.