GPN‑Star Model Predicts Genome‑Wide Functional Constraints
GPN‑Star is a new genomic language model that predicts functional constraints across the human genome by leveraging multispecies sequence alignments. Developed by a team including Ye, Benegas, Albors, Li, and Song, the model was trained on a large collection of aligned genomes, including 239 primate species, and incorporates evolutionary conservation signals to assess the impact of genetic…
Key points
- GPN‑Star predicts genome‑wide functional constraints using multispecies alignments
- It surpasses prior models like Nucleotide Transformer and AlphaGenome in accuracy
- The model is open‑source on Zenodo and supports variant effect prediction
In benchmark tests, GPN‑Star outperformed earlier models such as the Nucleotide Transformer and AlphaGenome, achieving higher precision in identifying constrained genomic elements and improving the accuracy of variant effect predictions. The authors report that the model can annotate over 3 billion base pairs of the human genome, providing a comprehensive constraint map that can be used to prioritize variants in clinical and research settings.
The release of GPN‑Star, available on Zenodo, offers the scientific community an open‑source tool that bridges deep learning and evolutionary biology. By enabling more reliable interpretation of non‑coding variants, it supports efforts in precision medicine and functional genomics research.
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Google launches Gemini 4 Argon model globally for Pro and Ultra subscribers with introductory API pricing · 32 src
- Meta releases Muse Spark Open, an 816B-parameter open-weight multimodal model with 42B active parameters and 1M-token context · 20 src
- Microsoft AI chief says in-house model beats OpenAI's using quarter of compute · 1 src
- NVIDIA releases Nemotron 3.5 Super VL and Bolt edge models for enterprise agents · 8 src
- Black Forest Labs releases FLUX 3 Image Klein, a 5.6B-Parameter Open-Weights model for text-to-image generation · 3 src
Comments
via GitHub Discussions