CHI's High Content Analysis East conference was a great success. The conference had overriding themes in data informatics, supervised and unsupervised learning and the use of predictive modeling.
Anne Carpenter shared a user friendly interface in Cell Profiler for an interactive semi supervised cell level learning tool. Neil Carragher of Astra Zeneca showed some really interesting work on predicting mechanism of action thru the use of learning on multivariate image descriptors. And CMU’s Robert Murphy presented the use of learning and modeling on protein patterns in the cell. All in all we, are really beginning to see the "high content" promise of the industry come to fruition. It is finally becoming more common to see research combining hundreds of variable readouts from the cells for extracting more sophisticated knowledge. Hopefully we will see these themes advancing at the upcoming HCA event in January, from the 11-15th at the Fairmont hotel in San Francisco,
Also of interest were several advances in managing High Content Screening Data. Present were Michael Sjaastad of Molecular Devices, Martin Daffertshofer of Perkin Elmer (Evotec), and Karol Kozak,, LMC-RISC, Institute for Biochemistry . They all presented promising offerings in managing HCS data. It was interesting to note that Thermo (Cellomics) did not introduce any new tools in managing their HCS data, perhaps telling of the great foresight that Mark Collins had, when originally designing Store which has stood the test of many years.