Unsupervised clustering algorithms for flow/mass cytometry data

Abstract

This Final Year Project report documents the process of using dimension reduction and unsupervised clustering methods for clustering similar group of cells and to automate the discovery of cell populations from data sets generated from mass cytometry. Also, it documents the process of developing a website to display the details of mass cytometry datasets. Traditionally, flow cytometry is used to analyse physical and chemical properties of cells by flowing a stream of fluid containing the cells through a detection device. Mass cytometry uses antibodies and rare earth elements to tag the cells which are then analysed by the mass spectrometer based on the time-of flight of these cells.