Dr. Chris Bakal, Team Leader at the Institute of Cancer Research in London, spoke at the Labcyte Genomics Symposium about the benefits of high-content screening for phenotypic profiling and functional genomics. High-content screening can be used to examine signaling between cancer cells and their microenvironment. Cancer cells undergo multiple cell shape changes in order to metastasize, move in and out of the bloodstream and colonize a new tissue. 90% of cancer patient deaths are due to metastasis and understanding how cancer cells change shape could provide new therapeutics avenues.
Dr. Chris Bakal, Team Leader at the Institute of Cancer Research in London, spoke at the Labcyte Genomics Symposium 2016 in Edinburgh about the benefits of high-content screening for phenotypic profiling and functional genomics.
Dr. Bakal describes the workflows employed for high-content phenotypic studies at the Institute of Cancer Research. RNAi is used to produce gene knock downs, the cells are then imaged and the images segmented to obtain morphological signatures of up to 600 different features, that capture the ‘shape’ of each cell. A complex picture of the cells with thousands of different RNAi and conditions is built up, which generates large data sets.
The large data sets are examined in network models that assess the architecture, dynamics and remodeling. Experimental validation is then carried out using the network models, with emphasis given to in vivo models. By knocking down genes, quantifying the resultant cell shapes and using different computational methods to cluster morphological signatures, Dr. Bakal’s group has described more than 35 different local signaling networks that regulate different aspects of cell shape. Dr. Bakal describes how it is possible to extrapolate protein function, location and protein-protein interactions of unknown RNAi targets from morphological analysis alone, which enables the characterization of hundreds of unknown genes.
Dr. Bakal gives examples of how his group applies high-content screening for ‘integrative image-omics’ to describe causal relationships between signaling network components. Dr. Bakal also describes the contribution of statistics and Bayesian inference and dependency modeling methods to build a picture of causality.
For several transcription factors, cell shape regulates signaling and not the other way around; this has important implications for screening. If cell shape is impacting signaling and therefore the phenotype, this must be taken into consideration when analyzing results from phenotypic studies. Dr Bakal describes a mathematical model to distinguish true ‘hits’ versus indirect effects, by correcting for cell shape changes in phenotypic screens. Problems with reproducibility between laboratories are common in RNAi screening and Dr. Bakal discusses how this could in part be due to failure to correct for these effects.