Yksittäisten solujen transkriptionaalisten nopeuksien mallintaminen
Year: 2020
Type: dissertation
Abstract: A recently developed measurement technique, called single-cell RNA sequencing (scRNA-seq), has provided us to study biological samples with unprecedented resolution, as we are now capable of studying tissues at single-cell level. As a result of scRNA-seq, new cell types have been discovered, detailed developmental paths of various tissues have been revealed, and regulatory relationships between genes have been uncovered. During a last few years, especially the research of differentiation trajectories of single cells has gained popularity, because it can broaden our knowledge of development of both normal and abnormal tissues. These trajectory inference studies can potentially let us understand better causes of various diseases and later on help us to discover new treatments against several pathogens. The aim of this thesis is to develop a new tool to analyse scRNA-seq data and especially the dynamics and the differentiation trajectories of the single cells. This method relies on a theory of RNA velocities discovered by La Manno et al. (2018), but has a unique approach to solve the latent dynamics: we utilise variational autoencoders (VAEs) to project the cells to the low-dimensional latent space and nonparametric ordinary differential equations (ODEs) to solve the latent differentiation trajectories. With our new method, we obtain rather promising results as the differentiation trajectories for two data sets are in agreement with the current knowledge of the development of the studied tissues. The novelty of our method is that we are capable of selecting any single cell from the data set and study its individual differentiation path or even intervene in its gene expression and follow the differentiation of the perturbed cell. These results suggest that our method has potential to become a widely applicable tool for analysing scRNA-seq data, but some aspects still require further refining.
Source: Aaltodoc (Aalto University)
Author: Vikkula, Salla
Institution: Aalto University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7113109901
