A new platform called the Perturbation Catalogue, brings together different types of gene perturbation data including CRISPR knockout screens, variant effect assays, and gene expression profiling. 

As genetic screening technologies advance, the number of datasets generated is growing rapidly but these can be siloed across different formats and repositories, making it difficult to find and compare these data effectively. The Perturbation Catalogue aims to address this by integrating perturbation data into a coherent, standardised framework to make access to these data easier to find, compare  and reuse, and to support the development of machine learning models in this field.

What does Perturbation Catalogue include? 

The Perturbation Catalogue is an Open Targets project, developed at EMBL-EBI in collaboration with the Wellcome Sanger Institute and Human Technopole. Researchers can search by gene, variant or phenotype, view summaries of experimental data, and explore harmonised metadata across datasets.

The beta release includes curated data from:

• MaveDB: multiplexed variant effect data
• DepMap: genome-wide CRISPR screens
• scPerturb and PerturBase: expression-level perturbation data

The platform is supported by Google BigQuery and Google Cloud infrastructure, using Pydantic – a data validation library for Python – to ensure data consistency and metadata standardisation. The Perturbation Catalogue will be integrated with other EMBL-EBI resources including Ensembl, UniProt, PDBe, and ChEMBL.

Try the beta version and send your feedback

The beta version of the Perturbation Catalogue is now live. Please explore the platform, test the APIs, and provide feedback on your experience. The platform will continue to grow and your input will help guide future developments.

Get access to the code and documentation on the Perturbation Catalogue GitHub repository. 

Please send your feedback to perturbation-catalogue-help@ebi.ac.uk or create GitHub issue.