Genomic prediction models such as GBLUP and its additive-dominance extension (GBLUP-AD) rely on predefined fixed genotype encodings, which implicitly assume symmetric allele effects and often fail to capture non-canonical gene action. We propose LatentGP, an end-to-end differentiable Gaussian process framework that learns genotype encoding as a model parameter and jointly estimates latent genotype representations and variance components through marginal log-likelihood maximization. Simulations demonstrated that LatentGP can recover complex gene action modes, particularly asymmetric super-additive patterns that were poorly captured by conventional models. In hybrid pig and maize populations, LatentGP achieved relative accuracy improvements ranging from 0.1% to 10.9% compared with baseline methods. Beyond predictive performance, visualization of the learned embeddings revealed asymmetric super-additive patterns underlying pig production traits and nonlinear dominance patterns for maize yield. In human disease cohorts, LatentGP automatically converged to a linear additive encoding, providing a negative control that highlights its ability to self-regulate according to the underlying genetic architecture. By shifting from fixed genotype coding to adaptive representation learning, LatentGP uncovers diverse global gene action patterns while preserving the statistical rigor of linear mixed models. The software is freely available at https://github.com/hyacz/LatentGP .
Zhang, Haohao, Tang, Zhenshuang, Lai, Chunxiao et al. Uncovering global gene action patterns via learnable genotype encoding. // Science China Life Sciences. – 2026. – DOI: 10.1007/s11427-026-3526-3. – URL: http://link.springer.com/openurl/fulltext?id=doi:10.1007/s11427-026-3526-3
Zhang, Haohao, Tang, Zhenshuang, Lai, Chunxiao, Peng, Jing, Fu, Yuhua, Xiong, Xiong, Zhou, Xiang, Cheng, Chong, Yin, Lilin, Yuan, & Xiaohui (2026). Uncovering global gene action patterns via learnable genotype encoding. Science China Life Sciences. https://doi.org/10.1007/s11427-026-3526-3
Zhang, et al. “Uncovering global gene action patterns via learnable genotype encoding.” Science China Life Sciences, 2026. doi:10.1007/s11427-026-3526-3.
@article{Zhang2026,
title = {Uncovering global gene action patterns via learnable genotype encoding},
author = {Zhang and Haohao and Tang and Zhenshuang and Lai and Chunxiao and Peng and Jing and Fu and Yuhua and Xiong and Xiong and Zhou and Xiang and Cheng and Chong and Yin and Lilin and Yuan and Xiaohui},
journal = {Science China Life Sciences},
year = {2026},
doi = {10.1007/s11427-026-3526-3},
url = {http://link.springer.com/openurl/fulltext?id=doi:10.1007/s11427-026-3526-3}
}
Traditional conservation practices, such as nature worship, have played an important role in protecting biodiversity in India. Sacred forests are well-preserved areas with high biodiversity. However, these forests vary in area and...
Freshwater fishes are very sensitive to changes in their environment, especially in fragile ecosystems of the Himalayas in their reproductive patterns. This is a comparative study of reproductive biology of three sympatric snowtrout...
This study investigated how PEG-induced osmotic stress affects germination, growth, morphology, physio-biochemical traits, and stress index parameters in 14 foxtail millet genotypes during early seedling stages. Twenty-five qualitative...
Insect herbivores face a risk-prone dilemma ‘to feed or to flee’ when there is an overlap in the microhabitat preferences of the potential prey and their arthropod predators. Field observations revealed the occurrence of five major...
The current research highlights the recollection of Solanum capsicoides All. in the Shi-Yomi district in Arunachal Pradesh, India, marking an extended distribution in this region. During a field survey in 2018 near Payum village in...