CRISPR screens: Data Collection and Prediction of Drug Sensitivity using Public Resources
Open AccessHigh-throughput screening has become a powerful and indispensable tool for systematic functional analysis of whole genomes. The volume of screening data grows massively, provides an abundance of information for functional genomics studies and drug target identification. We present CRISP-view, an integrated database of CRISPR/Cas9 and RNAi screening datasets with complete metadata annotation, covering multiple phenotypes and a variety of research areas. Currently, CRISP-view collects 11,384 in vivo or in vitro screening samples from 169 up-to-date publications. All the datasets are processed by a standardized MAGeCK-VISPR pipeline, allowing them to be compared across different experiments. A user-friendly web interface of CRISP-view is also available, which enables users to browser and search all the datasets, the metadata, and the quality control results. Additionally, we developed a machine learning model to predict the response of cell lines towards drug treatment, utilizing the beta scores of perturbed genes across cell lines and growth inhibitory activity with drugs. Our model got a precision of 0.889 and an overall F1 score, 0.391, suggesting that our model is useful and accurate. Moreover, we filter the crucial genes for each drug with high feature importance from the model, these genes are potential biomarkers for drug response. The BRAF and SOX10 gene with PLX could be potent evidence to support our results.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.