PRESCOTT: a population aware, epistatic and structural model accurately predicts missense effect

Predicting the functional impact of point mutations is a complex yet vital task in genomics. PRESCOTT stands at the forefront of this challenge and reconstructs complete mutational landscapes of proteins, enables the identification of protein regions most vulnerable to mutations and assigns scores to individual mutations, assisting pathologists in evaluating the pathogenic potential of missense variants. PRESCOTT categorizes these variants into three distinct classes: Benign, Pathogenic, or Variants of Uncertain Significance (VUS). The model leverages protein sequences across millions of species, advanced protein structural models, and extensive genomic and exomic data from diverse human populations. By using only sequence and structural information, it significantly improves on current standards for predicting mutations in human proteins and matches AlphaMissense performance, which incorporates allele frequency data in its analysis. By including population-specific allele frequencies, PRESCOTT excels in genome-scale score separation of ClinVar benign and pathogenic variants and surpasses AlphaMissense in analyzing the ACMG reference human dataset and the over 1800 proteins from the Human Protein Dataset. Its efficacy is particularly notable in autoinflammatory diseases, accurately predicting pathogenic gain-of-function missense mutations, a task known for its difficulty. Efficiency and accessibility are key aspects of PRESCOTT. The user-friendly PRESCOTT webserver facilitates mutation effect calculations on any protein and protein variants. The server hosts a Comprehensive Human Protein Database for over 19,000 human proteins, based on sequences and structures, ready for a customized allele population analysis. Additionally, the tool provides open access to all intermediate scores, ensuring interpretability and transparency in variant analysis. PRESCOTT is a significant stride forward in the field of genomic medicine, offering unparalleled insights into protein mutational impacts. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement French Agence Nationale de la Recherche (SolvingMEFVariants, ANR-21-CE17-0046) ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: (DMS data on human proteins) I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All data produced are available online at All the data generated in this work is provided in dedicated Zenodo repositories. ESCOTT predictions for the 3013 human proteins of EVE dataset: . ESCOTT predictions for the ProteinGym Dataset: . ESCOTT predictions for the Human Proteome dataset of more than 19,000 proteins: . iGEMME predictions of the Human Proteome dataset of more than 19,000 proteins: . The gain-of-function ESCOTT single point mutational data was extracted from the Human Proteome dataset in . The ACMG ESCOTT data was extracted from: .