Pathogenic variants underlying Mendelian diseases often disrupt the normal physiology of afew tissues and organs. However, variant effect prediction tools that aim to identifypathogenic variants are typically oblivious to tissue contexts. Here we report a machine-learning framework, denoted ‘Tissue Risk Assessment of Causality by Expression forvariants’ (TRACEvar, https://netbio.bgu.ac.il/TRACEvar/), that offers two advancements.First, TRACEvar predicts pathogenic variants that disrupt the normal physiology of specifictissues. This was achieved by creating 14 tissue-specific models that were trained on over14,000 variants and combined 84 attributes of genetic variants with 495 attributes derivedfrom tissue omics. TRACEvar outperformed 10 well-established and tissue-oblivious varianteffect prediction tools. Second, the resulting models are interpretable, thereby illuminatingvariants' mode-of-action. Application of TRACEvar to variants of 52 rare-disease patientshighlighted pathogenicity mechanisms and relevant disease processes. Lastly, interpretationof large-scale models revealed that top-ranking determinants of pathogenicity includedattributes of disease-affected tissues, particularly cellular process activities. Hence, tissuecontexts and interpretable machine-learning models can greatly enhance the etiology of rarediseases. Article link: https://www.embopress.org/doi/full/10.1038/s44320-024-00061-6

File Description full_dataset_2_stars_2022.csv training set new_test_set_2023_diff.csv test set full_dataset_2_stars_2023.csv full dataset Relevant_Columns_Names_Edited_2.csv Column names for tissue-specific models Transfer_Learning_Data_hg37.csv multi-tissue model data Transfer_Learninig_Relevant_Columns_Names_Edited.csv multi-tissue column names