EUVIMEDEUVIMED entdecken
Crossref

MIRAGE: a multimodal deep learning framework for interpretable risk assessment of high myopia from genetic and retinal imaging data

Yijun Zhou, Shuaishuai Gao, Tao Wang, Wei Dai, Kaicheng Ye, Bowei Ma, Yao Zhou, Shuyan Fan, Yuanyuan Gui, Wenhao Ma, Zhen Ji Chen, Shilai Xing und 4 weitere

Human Molecular Genetics · 09.06.2026 · 35 (11)

DOI: 10.1093/hmg/ddag040 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Human Molecular Genetics, online erschienen am 09.06.2026. Verfasst von Yijun Zhou und 15 weiteren Autor:innen. Beteiligte Einrichtung: School of Biomedical Engineering, Hainan University, No. 58 Renmin Rd, Haikou, Hainan, 570228, China. Hauptthemen laut MEDLINE: Deep Learning, Myopia, Retina. Die Arbeit nennt 48 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

High myopia (HM) is a complex condition influenced by both genetic and environmental factors, yet its early prediction and clinical intervention remain challenging due to heterogeneous progression patterns. To support early identification of individuals at risk for HM, we developed MIRAGE, a deep learning framework combining exome-wide genotypes with fundus images for personalized prediction. The model combines DeepExGRS for genetic risk modeling and a convolutional network for imaging, fused via a gating attention mechanism. Applied to a cohort of 1991 individuals, MIRAGE achieved high predictive accuracy (AUC = 0.963), outperforming unimodal models. Interpretation using Integrated Gradients and Shapley values identified key genes (e.g. GJD2, FGF1) and image regions (optic disc, macula) contributing to predictions. Further, deep learning–based genetic risk models (DeepcvGRS and DeeprvGRS) outperformed traditional polygenic risk scores, capturing nonlinear interactions and improving HM prediction. Interaction analysis using Shapley interaction scores revealed 314 significant gene–gene interactions, including biologically relevant pairs such as PRSS56–GLI3 and SEMA4D–FRY. Overall, MIRAGE offers a scalable, interpretable, and accurate approach for HM risk stratification, paving the way to early intervention guided by both genetics and retinal imaging.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningMyopiaRetinaHumansGenetic Risk ScoreRisk AssessmentGenetic Predisposition to DiseaseExomeFemaleGenotypeConvolutional Neural Networks

Stichwörter der Autor:innen

fundus imagegeneticshigh myopiamultimodalrisk prediction

Autor:innen und Institutionen

  • School of Biomedical Engineering, Hainan University, No. 58 Renmin Rd, Haikou, Hainan, 570228, China.
  • Hainan Institute of Real World Data, No. 32 Kangxiang Rd, Qionghai, Hainan, 571437, China.
  • State Key Laboratory of Eye Health, Eye Hospital, Wenzhou Medical University, No. 270 West Xueyuan Rd, Wenzhou, Zhejiang, 325027, China.
  • Oujiang Laboratory, Zhejiang Lab for Regenerative Medicine, Vision and Brain Health, No. 999 Jinshi Rd, Wenzhou, Zhejiang, 325000, China.
  • Key Laboratory of Laboratory Medicine, Ministry of Education, Institute of Genomic Medicine, Wenzhou Medical University, No. 1 Zhongxin Rd, Wenzhou, Zhejiang, 325035, China.
  • Institute of PSI Genomics, Research Center, No. 100 Luodong Rd, Wenzhou, Zhejiang, 325000, China.

ORCID-Profile: Jianzhong Su, Yinghao Yao

Zitieren

APA

Zhou, Y., Gao, S., Wang, T., Dai, W., Ye, K., Ma, B., Zhou, Y., Fan, S., Gui, Y., Ma, W., Chen, Z. J., Xing, S., Su, J., Qu, J., Yuan, J., & Yao, Y. (2026). MIRAGE: a multimodal deep learning framework for interpretable risk assessment of high myopia from genetic and retinal imaging data. Human Molecular Genetics, 35(11). https://doi.org/10.1093/hmg/ddag040

Vancouver

Zhou Y, Gao S, Wang T, Dai W, Ye K, Ma B, et al. MIRAGE: a multimodal deep learning framework for interpretable risk assessment of high myopia from genetic and retinal imaging data. Human Molecular Genetics. 2026;35(11). doi:10.1093/hmg/ddag040

BibTeX
@article{Zhou2026,
  author = {Zhou, Yijun and Gao, Shuaishuai and Wang, Tao and Dai, Wei and Ye, Kaicheng and Ma, Bowei and Zhou, Yao and Fan, Shuyan and Gui, Yuanyuan and Ma, Wenhao and Chen, Zhen Ji and Xing, Shilai and Su, Jianzhong and Qu, Jia and Yuan, Jian and Yao, Yinghao},
  title = {MIRAGE: a multimodal deep learning framework for interpretable risk assessment of high myopia from genetic and retinal imaging data},
  journal = {Human Molecular Genetics},
  year = {2026},
  volume = {35},
  number = {11},
  doi = {10.1093/hmg/ddag040}
}
Kostenlos starten →