Automatic identification of subcellular phenotypes on human cell arrays.

摘要: Light microscopic analysis of cell morphology provides a high-content readout function and protein localization. Cell arrays microwell transfection assays on cultured cells have made phenotype accessible to high-throughput experiments. Both the localization each in proteome effect RNAi knock-down individual genes can be assayed by manual inspection images. However, use morphological readouts for functional genomics requires fast automatic identification complex cellular phenotypes. Here, we present fully automated platform screening combining human live arrays, microscopy, machine-learning-based classification methods. Efficiency this is demonstrated eleven subcellular patterns marked GFP-tagged proteins. Our method adapted virtually any assay based morphology, opening wide range applications including large-scale cells.