Code and models associated with the manuscrupt "Protein Condensate Atlas from predictive models of heteromolecular condensate composition".
The code enables to prepare the relevant datasets, perform comparisons between the datasets as outlined in the manuscript, train models for protein localisation into condensates and, finally, predict the Protein Condensate Atlas and perform the validation as has been outlined in the manuscript.
All the files in the folders 'Data' and 'Output' are accessible via doi 10.5281/zenodo.10207248.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
