![]() To build Gaussian mixture (GM) profiles, 10% of the data were subsampled uniformly across all samples. Our test experiment uses a physiologically relevant p53–wild-type breast cancer model system (MCF-7) and a mechanistically distinct set of targeted and cancer-relevant cytotoxic compounds that induces a broad range of gross and subtle phenotypes. We extracted the core profiling methods-namely, the algorithms for constructing per-sample profiles from per-cell measurements-from the larger methodologies, applied them to a typical experiment, and compared their ability to classify compounds into MOA. Little is known about how the methods compare because each method was proposed as part of a more extensive methodology, often with different goals and with different types of data available (multiple concentrations, cell lines, or marker sets). The methods range from simple and fast to complicated and computationally intensive, and they differ greatly in how explicitly they take advantage of the individual-cell measurements to describe heterogeneous populations. ![]() This article describes and compares five methods that have been proposed for profiling and shown to be effective in a particular experiment. Most image-based profiling experiments thus far have been performed at the proof-of-principle scale, with a focus on developing computational methods for generating and comparing profiles. To facilitate the ready application and future development of image-based phenotypic profiling methods, we provide our complete ground-truth and test data sets, as well as open-source implementations of the various methods in a common software framework. Another simple method, which performs factor analysis on the cellular measurements before averaging them, provided substantial improvement and was able to predict MOA correctly for 94% of the treatments in our ground-truth set. This is surprising because many treatments induced a heterogeneous phenotypic response across the cell population in each sample. A very simple method that is based on population means performed as well as methods designed to take advantage of the measurements of individual cells. We compared these methods by applying them to a widely applicable assay of cultured cells and measuring the ability of each method to predict the MOA of a compendium of drugs. Several profiling methods have been proposed, but little is known about their comparative performance, impeding the wider adoption and further development of image-based profiling. Recently, image-based profiling has shown promise as a means for broadly characterizing molecules’ effects on cells in several drug-discovery applications, including target-agnostic screening and predicting a compound’s mechanism of action (MOA). Quantitative microscopy has proven a versatile and powerful phenotypic screening technique.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |