D2Deep: Combining evolution and protein language models for cancer driver mutation prediction

Datasets containing predictions, training and validation data for D2Deep predictor:

• D2Deep_predictions: D2Deep predictions for mutations in cancer driver proteins included in Next Generation Sequencing (NGS) panel of biopsies of haematological and solid tumours from Compermed Guidelines (https://www.compermed.be/en/guidelines)
• common_variants: common variants from gnomAD database (December 2022)
• dbSNP: Single nucleotide polymorphisms (SNPs) from the Single Nucleotide Polymorphism database (dbSNP)
• humsavar_benign_mutations: UniProtKB/Swiss-Prot human missense variants - release 21st December 2021
• clinvar_benign_deleterious_missense: ClinVar missense variants (March 2023)
• Tier.csv: Missense Tier 1,2,3 mutations from Catalogue of Somatic Mutations in Cancer (COSMIC - Cancer Mutation Census releasev92)
• cgi.csv: Missense oncogenic mutations from Cancer Genome Interpreter (release 2018)
• Balanced_training_set: Pathogenic/benign balanced set (on gene level) used for training the model
• log_probWT_MUT_Tier1_2_3_common_balanced+-2_2200AA_57maxpool: Training set features used for model training
• DMS_mutations: Deep Mutational Scanning mutations used for validation (2021 - https://doi.org/10.15252/msb.202110305)
• DRGN_testset: DRGN test set used for validation
• clinvar_balanced_somatic_germline_missense: Clinvar somatic versus germline subset used for validation (March 2023) 
• 5genes_clinvarlabels_D2D_confidence_all: Performances of 6 predictors on 5 cancer genes mutations (March 2023)
• TP53_expert_multiple_single_submitters, BRAF_expert_multiple_single_submitters, CHEK2_expert_multiple_single_submitters, AR_expert_multiple_single_submitters, PTEN_expert_multiple_single_submitters : ClinVar labels with Review status: Practice guideline, Expert panel, Multiple submitters, Single submitter (March 2023)
• all_msas: mmseq2 Multiple Sequence Alignments for proteins used

                                ------------------------------------------------------------------------------------------------------------------------

                                               You can use our web server to query protein mutations and use the

                                                       interactive visualizations: https://tumorscope.be/d2deep/

                                ------------------------------------------------------------------------------------------------------------------------