White Rose University Consortium logo
University of Leeds logo University of Sheffield logo York University logo

Statistical Analysis of High Content Screening Data

Jacques, Richard (2009) Statistical Analysis of High Content Screening Data. PhD thesis, University of Sheffield.

[img]
Preview
Text
Jacques,_Richard.pdf
Available under License Creative Commons Attribution-Noncommercial-No Derivative Works 2.0 UK: England & Wales.

Download (941Kb)

Abstract

High throughput screening experiments are typically used within the pharmaceutical industry for the identification and evaluation of candidate drugs. Using a high throughput screen with automated imaging platform allows a large number of compounds to be tested in a biological assay in order to identify any activity inhibiting or activating a biological process. High throughput fluorescent images contain information that can be used to define fully the effects of a compound on cells. It is for this reason that florescent imaging assays have been termed high content screening (Clemons, 2004). The studies analysed in this thesis involve the use of an automated robotic system to administer compounds to cellular assays and take high content images. These images are then analysed and quantified using imaging algorithms to produce a set of variables. Each high content screen may extend to a million or more individual assays. Supervised classification methods have important applications in high content screening experiments where they are used to predict which compounds have the potential to be developed into new drugs. The use of supervised classification for high content screening data is investigated and a new classification method is proposed for batches of compounds where the rule is updated sequentially using information from the classification of previous batches. This methodology accounts for the possibility that the training data are not a representative sample of the test data and that the underlying group distributions may change as new compounds are analysed. Unsupervised classification methods are used in the analysis of high content screening experiments to evaluate potential new drugs. The study in this thesis considers clustering compounds based on their toxicological effect on the liver. Drug induced liver injury is the most common cause for non approval and withdrawal by the Food and Drug Administration (Ainscow, 2007a) and therefore this is an important stage in drug development.

Item Type: Thesis (PhD)
Academic Units: The University of Sheffield > Faculty of Science (Sheffield) > School of Mathematics and Statistics (Sheffield)
Depositing User: Dr Richard Jacques
Date Deposited: 19 Mar 2012 10:04
Last Modified: 08 Aug 2013 08:48
URI: http://etheses.whiterose.ac.uk/id/eprint/2220

Actions (repository staff only: login required)