On the Application of Active Learning and Gaussian Processes in Postcryopreservation Cell Membrane Integrity Experiments

Authors:
Mindaugas Norkus;Damien Fay;Mary J. Murphy;Frank Barry;Gearoid OLaighin;Liam Kilmartin
Affiliations:
National University of Ireland, Galway;University of Cambridge, Cambridge;National University of Ireland, Galway;National University of Ireland, Galway;National University of Ireland, Galway;National University of Ireland, Galway
Venue:
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Year:
2012

Citing 8
Cited 0

Information-based objective functions for active data selection

Neural Computation
Regression with input-dependent noise: a Gaussian process treatment

NIPS '97 Proceedings of the 1997 conference on Advances in neural information processing systems 10
Gaussian Process Regression: Active Data Selection and Test Point Rejection

Mustererkennung 2000, 22. DAGM-Symposium
Bayesian treed gaussian process models

Bayesian treed gaussian process models
Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning)

Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning)
Most likely heteroscedastic Gaussian process regression

Proceedings of the 24th international conference on Machine learning
Towards Real-Time Information Processing of Sensor Network Data Using Computationally Efficient Multi-output Gaussian Processes

IPSN '08 Proceedings of the 7th international conference on Information processing in sensor networks
Near-Optimal Sensor Placements in Gaussian Processes: Theory, Efficient Algorithms and Empirical Studies

The Journal of Machine Learning Research

Quantified Score

Hi-index	0.00

Visualization

Abstract

Biological cell cryopreservation permits storage of specimens for future use. Stem cell cryostorage in particular is fast becoming a broadly spread practice due to their potential for use in regenerative medicine. For the optimal cryopreservation process, ultralow temperatures are needed. However, elevated temperatures are often unavoidable in a typical sample handling cycle which in turn negatively affects the postcryopreservation integrity of cells. In this paper, we present an application of active learning using an underlying Gaussian Process (GP) model in an experimental study on postcryopreservation membrane integrity response to a range of elevated temperature conditions. We tailored this technique for the current investigation and developed an algorithm which enabled identification of the sampling locations for the experiments in order to obtain the highest information return about the process from a limited size sample set. We applied this algorithm in the experimental study investigating the effects of severe temperature elevation (ranging from -40 to 20^{\circ }{\rm C}) over a short term event (48 hours) on the postcryopreservation membrane integrity of Mesenchymal Stem Cells (MSCs) derived from human bone marrow. The algorithm showed excellent performance by selecting the locations which maximized the reduction of variance of the process response estimate. An approximating GP model developed from this experimental data shows that the elevated temperatures during cryopreservation have an imminent detrimental effect on cell integrity.