![]() Detecting and removing outlier cells can highly improve the profile quality. For instance, an error in the segmentation step may result in overly small or large cells and bias the profiles heavily as a result. Outlier cells, which do not show any valid biological effect, may result from errors in different parts of the pipeline. A significant step in image-based profiling and data cleaning is cell-level quality control. Hence, in this work, we focus on extracted image-based profiles as the input to the downstream analysis. However, examining raw images can be time-consuming and not applicable to previously prepared datasets. There are some quality control methods applied to raw images for instance, in, a cell-level quality control approach has been proposed based on the high throughput images. Data cleaning is a key step for enhancing image-based profiling as there may be different artifacts in the staining and imaging process and can affect the next steps. We conduct a comprehensive study on how such techniques could potentially improve the profile’s quality. We instead mainly focus on approaches that preprocess the features extracted by CellProfiler, open-source software that aims to automate most of these steps. The initial steps such as the illumination correction, segmentation, and feature extraction are not investigated in this work. The typical workflow in the analysis of images that are produced by high-throughput assays includes illumination correction, nuclei/cell segmentation, quality control, morphological feature measurement, batch effect removal, data normalization, feature selection/dimensionality reduction, and finally, aggregation of single cell measurements into image-based profiles per well. Therefore, high throughput assays can be helpful in this process. According to previous studies, the MoA of unknown compounds can be predicted by grouping each unknown compound with already-annotated compounds based on the similarity of their morphological profiles. Prediction of drug MoAs through such assays potentially saves drug discovery process costs when applied early on. ![]() Image-based profiling has diverse and powerful applications, including identification of gene and allele functions and targets, and mechanisms of action (MoA) of drugs. These experiments often involve growing cells in multi-well plates and then treating cells in each well with a small molecule, or genetic perturbation. High-throughput image-based assays have proved to be an effective predictive tool in the early stages of drug discovery through automated microscopy and image analysis, which make quantification of cellular morphological responses possible at a large scale. In the end, we also suggest possible avenues for future research. Our experiments indicate that by performing these time-efficient preprocessing steps, image-based profiles can preserve more meaningful information compared to raw profiles. Our enhancement steps mainly consist of data cleaning, cell level outlier detection, toxic drug detection, and regressing out the cell area from all other features, as many of them are widely affected by the cell area. We consider the identification of drug mechanisms of action as the downstream task to evaluate such preprocessing approaches. In this work, we examined various preprocessing approaches to improve the profiles. However, there may be several sources of error in the CellProfiler quantification pipeline that affects the downstream analysis that is performed on the profiles. Single cell features are then aggregated for each treatment replica to form treatment “profiles”. CellProfiler is a popular and commonly used tool for this purpose through providing readily available modules for the cell/nuclei segmentation, and making various measurements, or features, for each cell/nuclei. Image analysis pipelines have a pivotal role in translating raw images that are captured in such assays into useful and compact representation, also known as measurements. Indeed, these assays have proved to be effective in characterizing unknown functions of genes and small molecules. Image-based assays are among the most accessible and inexpensive technologies for this purpose. With the advent of high-throughput assays, a large number of biological experiments can be carried out.
0 Comments
Leave a Reply. |