Research map / Architectures and efficiency / Attention and transformers

PRSformer: Disease Prediction from Million-Scale Individual Genotypes

NeurIPS 2025 · Architectures and efficiency › Attention and transformers · approach: Transformer neural operators for PDEs · 1 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces PRSformer, a scalable deep learning architecture for disease risk prediction from million-scale individual-level genotypes, aiming to capture genetic interactions beyond linear summary-statistic polygenic risk scores.

Problem

Predicting disease risk from individual genotypes at biobank scale (N>10^6 individuals, L>10^5 genotype features), where current methods are often linear and rely on summary statistics, missing complex interactions and individual-level information.

Approach

A scalable transformer-based deep learning model applied directly to individual-level genotypes to capture complex genetic interactions for polygenic risk prediction, in contrast to linear summary-statistic PRS methods.

Closest papers by meaning

• Training Flexible Models of Genetic Variant Effects from Functional Annotations using Accelerated Linear Algebra (ICML 2025 · similarity 0.85)
• High-Dimensional Sensitivity Analysis for Genomic Studies: An Adversarial Framework for Learning Worst-Case Latent Confounders (ICML 2026 · similarity 0.84)
• GV-Rep: A Large-Scale Dataset for Genetic Variant Representation Learning (NeurIPS 2024 · similarity 0.84)
• GeSubNet: Gene Interaction Inference for Disease Subtype Network Generation (ICLR 2025 · similarity 0.83)
• Learning Identifiable Factorized Causal Representations of Cellular Responses (NeurIPS 2024 · similarity 0.82)
• PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (NeurIPS 2025 · similarity 0.82)