The findings focus on translating genetic changes into disease pathways to inform diagnostics and treatment strategies, advancing the clinical utility of genotype-to-phenotype predictions.

Key quotes from first author David Stein and senior author Dr. Avner Schlessinger emphasize faster, more accurate, and clinically relevant results for rare diseases.

The study appears in Nature Communications on December 15, 2025, in an article titled “Expanding the utility of variant effect predictions with phenotype-specific models,” with authors including David Stein, Meltem Ece Kars, Baptiste Milisavljevic, Matthew Mort, Peter D. Stenson, Jean-Laurent Casanova, David N. Cooper, Bertrand Boisson, Peng Zhang, Avner Schlessinger, and Yuval Itan.

Funding for the research comes from NIH grants (including R24AI167802, P01AI186771, R01CA277794, R01HD107528, R01NS145483, S10 grants) and foundations such as Fondation Leducq and the Helmsley Charitable Trust, underscoring strong financial backing.

Senior authors say V2P could streamline diagnosis, shorten diagnostic odysseys for unexplained conditions, and guide targeted research and treatment strategies.

The study demonstrates V2P’s potential for precision medicine and rare-disease research, showcased in a Nature Communications online publication.

The model currently classifies variants into broad disease categories, with researchers aiming to refine predictions toward specific diseases and to integrate data on gene interactions and protein function.

Contributors to the study include Meltem Ece Kars, Baptiste Milisavljevic, Matthew Mort, Peter D. Stenson, Jean-Laurent Casanova, David N Cooper, Bertrand Boisson, and Peng Zhang.

The work is published in the December 15 online issue of Nature Communications.

A new AI tool called V2P (Variant to Phenotype) from Icahn School of Medicine at Mount Sinai links genetic mutations directly to the diseases they are likely to cause, potentially enabling earlier and more precise diagnoses.

Trained on a large database of variants with known outcomes and tested on de-identified patient data, V2P often ranks the true disease-causing variant among the top ten candidates.

The paper lists contributing authors and notes that David Stein and Yuval Itan are among those shaping the work.