Abstract

SOM Analysis of individual cellular fingerprints in High Content Images

Martin Daffertshofer, Director Strategic Marketing Imaging Informatics, Perkin Elmer Cellular Technologies GmbH

High Content Instrumentation nowadays captures 1.000.000 images per day with subcellular resolution (40x, 60x, …) each showing about 500 individual cells. In this presentation the challenges analyzing vast amounts of cells each characterized by a rich set of morphological, texture and fluorescence intensity parameters will be introduced. A concept on how to utilize High Performance Computing in a Cloud environment will be described to allow artificial intelligence base grouping of cell populations using either supervised or unsupervised approaches. The relation of cell populations with chemical structure information to support small molecule screening as well as Gene Expression Data to support siRNA based functional screening benefits by the open architecture in the Cloud based system. Results will be discussed in the context of the image analysis quality on the one hand and with respect to the challenges in terms of the visualization of massive amounts of scientific data.