An end-to-end framework for the prediction of protein structure and fitness from single sequence

Info Feature, model parameters and prediciton results for BioRixv manuscript titled "An end-to-end framework for the prediction of protein structure and fitness from single sequence". For source code and detailed instructions on usage, please refer to our GitHub. If you want to quickly use our tool, you can utilize our web server for SPIRED-Fitness and SPIRED-Stab. Model Note: Only model parameters are available. For detailed instructions on usage, please refer to our GitHub. model.zip SPIRED-Fitness.pth SPIRED-Stab.pth SPIRED Predicted Structure CAMEO.tar.gz Inference results for CAMEO targets (August 2022∼August 2023), consisting of 680 protein chains with the length ranging from50 to 1,126 residues. Including predicted and label PDB files and tables for TM-score, lDDT & RMSD information. CASP15.tar.gz Inference results for 45 target domains from the CASP15 competition. Including predicted and label PDB files and tables for TM-score, lDDT & RMSD information. SCOPe.tar.gz Inference results for 1,231 folds (34,021 domains) from SCOPe database (v2.08, S95, September 2021). Including predicted PDB files and tables for TM-score, lDDT & RMSD information. CATH.tar.gz Inference results for 1,223 topologies (24,183 domains) from CATH database (v4.2, S35, July 2017). Including predicted PDB files and tables for TM-score, lDDT & RMSD information. Fitness Prediction fitness_data.zip single_and_double_mutation_fitness_train_data.csv single_mutation_fitness_pred.csv Stability Prediction stability_data.zip ddG_S669_test_data.csv ddG_S461_test_data.csv ddG_train_data.csv dTm_S557_test_data.csv dTm_train_data.csv SPIRED-Fitness Stage 2 Training/Validation Data Example Here, we provide SPIRED-Fitness Stage 2 training/validation data and the parameters of the stage 1 SPIRED-Fitness (SPIRED.pth and Fitness.pth). Note that due to size constraints, only structure data of training/validation samples are provided. If you need the fitness data of training/validation samples, please contact authors. For the complete training code, please visit our GitHub repository. SPIRED-Fitness_Stage2_training_data.zip SPIRED-Stab Stage 1 features wt_data_for_value_function.pt FoldX Predicted Structure cDNA_dataset_FoldX_structure.zip Mutant structure (PDB file) of each entry in cDNA proteolysis dataset generated by FoldX