Huang Lei, Su Kuan-Jui, Song Meng, Qiu Chuan, Gragert Loren, Deng Jeffrey, Luo Zhe, Tian Qing, Gong Ping, Shen Hui, Zhang Chaoyang, Deng Hong-Wen
University of Southern Mississippi.
Tulane University.
Res Sq. 2025 Aug 26:rs.3.rs-6797863. doi: 10.21203/rs.3.rs-6797863/v1.
Missing genotypes reduce statistical power and hinder genome-wide association studies. While reference-based methods are popular, they struggle in complex regions and under population mismatch. Existing reference-free deep learning models show promise in addressing this issue but often fail to impute rare variants in small datasets. We propose BiU-Net, a biologically informed U-Net model that segments genotype data and encodes positional information to preserve the genomic context. Evaluated on the 1000 Genomes Project, Louisiana Osteoporosis Study, and Simons Genome Diversity Project datasets, BiU-Net outperformed Beagle and sparse convolutional denoising autoencoder in overall metrics and in metrics stratified by minor allele frequency.
缺失的基因型会降低统计效力并阻碍全基因组关联研究。虽然基于参考的方法很流行，但它们在复杂区域和群体不匹配的情况下会遇到困难。现有的无参考深度学习模型在解决这个问题方面显示出了前景，但在小数据集中往往无法估算罕见变异。我们提出了BiU-Net，这是一种具有生物学信息的U-Net模型，它对基因型数据进行分割并编码位置信息以保留基因组背景。在千人基因组计划、路易斯安那骨质疏松症研究和西蒙斯基因组多样性计划数据集上进行评估时，BiU-Net在总体指标以及按次要等位基因频率分层的指标方面均优于Beagle和稀疏卷积去噪自动编码器。
