Multistage Optimal Transport evaluated on scRNA-seq Human Bone Marrow data

Title:  Multistage Optimal Transport evaluated on scRNA-seq Human Bone Marrow data
Author: Vilhelm Conradi & Hugo Nelsson
Date & Time: June 4th, 11:00–11:30
Location: Seminar Room M 3170-73 in the M-building, LTH
Supervisor: Anders Rantzer and Karl Rydén
Examiner:  Magnus Fontes


Abstract:

With novel techniques like single cell RNA sequencing we can look inside single cells in great detail and count the amounts of individual genes present in them. When modeling differentiation of cells one typically would use longitudinal data but assuming that every cell state in the differentiation process is present in one cell sample we presume that we have sufficient information to model the differentiation process from snapshot data. Building on the assumption that cell states similar to each other are closer in the space defined by their genes, and that nature would take the path through the gene space with the smallest cost, we use a newly proposed optimal transport algorithm to find differentiation trajectories through bone marrow samples and order cell adhering to Hematopoiesis in pseudotime. We then extract lists of genes describing these trajectories and evaluate their function to asses the viability of the new optimal transport algorithm. The ambition is that this method can aid the study of the development of Acute Myeloid Leukemia