B1015
Title: Bayesian approach for modelling RNA velocity
Authors: Elena Sabbioni - Politecnico di Torino (Italy) [presenting]
Enrico Bibbona - Politecnico di Torino (Italy)
Gianluca Mastrantonio - Politecnico of Turin (Italy)
Guido Sanguinetti - Scuola Internazionale di Studi Superiori Avanzati - SISSA (Italy)
Abstract: RNA velocity is a biological concept used to interpret cellular differentiation, exploiting the evolution of gene expression of cells at different stages of maturity during the evolution path. Gene expression is regulated by processes such as transcription, splicing, and degradation, which can be modeled as a stochastic chemical reaction network, where the rate constants need to be estimated using experimental data obtained from single-cell RNA sequencing. The primary method used in RNA velocity to estimate the parameters of the model is called scVelo. However, this approach has faced criticisms on various fronts. To address these limitations, a simplified model is proposed that is mathematically better founded and ensures the identifiability of parameters, unlike scVelo. The approach eliminates artificial preprocessing steps and instead employs a data distribution that is motivated by reaction network theory. Furthermore, cells of the same type are assumed to follow common dynamics.
