基于多源遗传信息的临床疾病风险评估系统

宁开达; 余正阳; 赵鑫; 李梓妍; 代菊; 夏立

doi:10.11999/JEIT251025

基于多源遗传信息的临床疾病风险评估系统

doi: 10.11999/JEIT251025 cstr: 32379.14.JEIT251025

宁开达¹,
余正阳²,
赵鑫^{1, 3},
李梓妍²,
代菊¹,
夏立^2, ,

1.
鹏城实验室深圳 518055
2.
华南理工大学数学学院广州 510640
3.
北京航空航天大学计算机学院北京 100191

基金项目: 鹏城实验室重大攻关项目(PCL2025AS212-3，PCL2024A02-2)，国家自然科学基金(12571529)，广东省基础与应用基础研究基金会(2024A1515-010699, 2022A1515-011426)

详细信息

作者简介:
宁开达：女，助理研究员，研究方向为医疗大数据建模、计算医学等

余正阳：男，硕士，研究方向为医学人工智能、生物信息学等

赵鑫：男，博士生，研究方向为医学图像模式识别、人工智能等

李梓妍：女，硕士生，研究方向为基因数据建模

代菊：女，助理研究员，研究方向为医疗大数据、图像模式识别等

夏立：男，教授，研究方向为多模态大数据建模

通讯作者:
夏立　lcxia@scut.edu.cn

中图分类号: TN911.7; TP391
计量
- 文章访问数: 98
- HTML全文浏览量: 35
- PDF下载量: 17
- 被引次数: 0
出版历程
- 收稿日期: 2025-09-28
- 修回日期: 2026-03-09
- 录用日期: 2026-03-09
- 网络出版日期: 2026-03-12
- 刊出日期: 2026-03-10

Clinical Disease Risk Assessment System Based on Multi-source Genetic Information

NING Kaida¹,
YU Zhengyang²,
ZHAO Xin^{1, 3},
LI Ziyan²,
DAI Ju¹,
XIA Li^{2
, ,}

1.
Pengcheng Laboratory, Shenzhen 518055, China
2.
School of Mathematics, South China University of Technology, Guangzhou 510640, China
3.
School of Computer Science and Engineering Beihang University School of Computer Science and Engineering, Beijing 100191, China

Funds: The Major Key Project of Pengcheng Laboratory (PCL2025AS212-3, PCL2024A02-2), The National Natural Science Foundation of China (12571529) , Guangdong Basic and Applied Basic Research Foundation (2024A1515-010699, 2022A1515011426)

摘要

摘要: 复杂疾病由多基因遗传与环境因素共同作用，致病机制高度异质，且具有高流行率和高致死率，构成重大公共卫生挑战。传统单疾病多基因风险评分(PRS)仅整合单一性状的遗传变异，忽视跨性状遗传相关性，预测效能受限；同时，多数方法依赖线性建模，难以刻画单核苷酸多态性(SNP)间及SNP与PRS之间的非线性交互，也未充分挖掘多类疾病PRS所蕴含的共享遗传信息。针对上述不足，该文提出一种基于统计学习的多源PRS疾病预测模型mtSNPPRS_XGB，创新性地构建SNP-PRS融合架构，利用XGBoost捕捉多源遗传特征的非线性交互。该模型联合整合原始SNP数据与多源PRS信息，在UK Biobank的18种疾病中取得平均AUC 66.70%(95%置信区间：66.46%~66.95%)，较传统UniPRS提升4.39%，较基于弹性网络的模型提升1.04%。研究为复杂疾病个体化遗传风险预测提供了新思路。
- 复杂疾病风险预测 /
- 多源PRS /
- SNP特征 /
- 非线性建模
Abstract: Objective Complex diseases are driven by polygenic inheritance and gene–environment interactions, resulting in highly heterogeneous pathogenic mechanisms and posing major challenges for both research and public health. Conventional single-trait polygenic risk scores (PRS) aggregate genetic variants associated with individual diseases but are limited by their neglect of cross-trait genetic correlations and nonlinear genetic interactions. Although multi-trait PRS approaches have been proposed to improve prediction accuracy, existing statistical-learning frameworks predominantly rely on linear integration of PRS features, failing to capture nonlinear interactions among single-nucleotide polymorphisms (SNPs) and to fully exploit shared genetic information across diseases. To address these limitations, we propose a nonlinear multi-source disease prediction framework, the SNP–PRS Fusion model, termed the mtSNPPRS_XGB (mtSNP-PRS XGBoost Integration Model). Methods The mtSNPPRS_XGB framework integrates raw SNP data of target traits with multi-trait PRS information to enhance genetic risk prediction for complex diseases through nonlinear modeling. SNPs significantly associated with target diseases were extracted from the GWAS Catalog (p < 5 × 10^–8) and encoded as allele dosages (0/1/2), while PRS weights covering 80 traits were obtained from the PGS Catalog and used to compute individual PRS. After standardized preprocessing, SNP and PRS features were jointly fused and modeled using XGBoost to capture complex SNP-SNP and SNP-PRS interactions. This framework introduces two key innovations: (1) collaborative modeling of multi-trait genetic information by jointly leveraging disease-specific SNPs and cross-disease PRS, and (2) systematic learning of nonlinear genetic interactions to overcome the linear constraints of conventional PRS-based models. Results and Discussions The mtSNPPRS_XGB model was evaluated using UK Biobank data across 18 complex diseases. It achieved an average AUC of 66.70%, representing improvements of 1.04% over the elastic-net-based model and 4.39% over the conventional UniPRS model. The inclusion of SNP features substantially improved predictive performance in diseases such as coronary heart disease, psoriasis, and celiac disease, while the integration of multi-trait PRS further enhanced specificity, particularly in cardiovascular, autoimmune, and cancer-related conditions. SHAP-based interpretability analyses demonstrated that mtSNPPRS_XGB simultaneously captures global cross-disease genetic liability encoded by PRS and disease-specific localized SNP effects, as illustrated in Alzheimer’s disease, colorectal cancer, gout, and ischemic stroke. These findings support both the biological plausibility and interpretability of the proposed framework. Conclusions We present a novel statistical learning-based multi-trait genetic risk prediction model, mtSNPPRS_XGB, which introduces an SNP-PRS fusion architecture and employs XGBoost to capture nonlinear interactions among multi-source genetic features. By integrating raw SNP data with multi-trait PRS, the proposed framework significantly improves risk prediction performance for complex diseases. Validation across 18 diseases in the UK Biobank demonstrates consistent performance gains over traditional PRS-based methods. This study overcomes the linear modeling limitations of conventional PRS approaches and provides a new paradigm for nonlinear integration of SNPs and multi-trait PRS, offering a robust and interpretable tool for personalized genetic risk prediction in precision medicine.
- Complex disease risk prediction /
- Multi-trait PRS /
- SNP characteristics /
- Nonlinear modeling

HTML全文

图 1 mtSNPPRS_XGB模型的创新框架

注：在SNP特征提取阶段，从GWAS Catalog中选取与目标性状显著相关的SNP位点；在PRS特征提取阶段，利用PGS Catalog计算多个相关性状的多基因风险评分(PRS)，以整合不同性状间的共享遗传信息；在分类决策阶段，将SNP特征与PRS特征进行融合拼接，并采用XGBoost模型对组合后的特征进行全局建模，以提升遗传风险预测的准确性。

下载: 全尺寸图片幻灯片

图 2 4种方法在18种疾病下的AUC性能对比

下载: 全尺寸图片幻灯片

图 3 4种方法在18种疾病上的预测性能对比

下载: 全尺寸图片幻灯片

图 4 各模型在6种疾病的性能对比

下载: 全尺寸图片幻灯片

图 5 mtSNPPRS_XGB模型在6种疾病的DCA性能分析

下载: 全尺寸图片幻灯片

图 6 4种疾病模型中SHAP最高的15个特征的蜂群图

每个点代表一个样本，横轴表示SHAP值，纵轴表示特征名称。点的颜色从蓝到红表示特征值从低到高，颜色越红表示该特征值越大。点的分布越分散表示该特征对不同样本的影响差异越大。SHAP值为正表示该特征增加疾病风险，为负则表示降低疾病风险。

下载: 全尺寸图片幻灯片

表 1 18种疾病的基本信息

疾病名称	缩写	中文名	ICD 10	正样本数	SNP数量
Coronary Artery Disease	CAD	冠心病	I21,I22,I23, I24.1,I25.2	39,523	1548
Heart Failure	HF	心衰	I50	14,411	173
Ischaemic Stroke	ISS	缺血性卒中	I63,I64	12,165	159
Alzheimer’s Disease	AD	阿尔兹海默病	F00,G30	3,077	63
Parkinson’s Disease	PD	帕金森病	G20	3,145	355
Bipolar Disorder	BD	双相情感障碍	F31	1,689	862
Breast Cancer	BC	乳腺癌	C50	12,694	884
Colorectal Carcinoma	CRC	结直肠癌	C18,C19,C20	7,027	44
Skin Cancer	SKC	皮肤癌	C44	19,143	538
Rheumatoid Arthritis	RA	类风湿关节炎	M05,M06, M08.0	8,835	1888
Psoriasis	PSO	银屑病	L40	10,865	515
Systemic Lupus Erythematosus	SLE	系统性红斑狼疮	M32	650	784
Gout	GO	痛风	M10	14,220	134
Celiac Disease	CED	乳糜泻	K90.0	2,455	153
Asthma	AST	哮喘	J45,J46	49,477	1241
Type 2 Diabetes Mellitus	T2D	2型糖尿病	E11	29,024	2131
Non-alcoholic Fatty Liver Disease	NAFL	非酒精性脂肪肝	K76.0	4,931	205
Glaucoma	GLAU	青光眼	H40	14,143	701

下载: 导出CSV

表 2 不同类别疾病下的各方法AUC性能对比(%)

疾病类别	方法	平均AUC	95% Lower CI	95% Upper CI
心血管疾病	mtPRS_XGB	64.60	64.50	64.70
	mtPRS_ML	64.83	64.76	64.90
	UniPRS	58.41	58.41	58.41
	mtSNPPRS_XGB	64.90	64.79	64.99
自身免疫疾病	mtPRS_XGB	65.30	64.96	65.63
	mtPRS_ML	66.07	65.83	66.33
	UniPRS	63.14	63.14	63.14
	mtSNPPRS_XGB	68.89	68.53	69.23
癌症	mtPRS_XGB	61.52	61.35	61.67
	mtPRS_ML	61.91	61.82	62.01
	UniPRS	61.43	61.42	61.44
	mtSNPPRS_XGB	62.77	62.60	62.94
精神类疾病	mtPRS_XGB	62.91	62.60	63.22
	mtPRS_ML	63.67	63.50	63.81
	UniPRS	62.93	62.93	62.93
	mtSNPPRS_XGB	63.51	63.12	63.97
其他疾病	mtPRS_XGB	69.76	69.71	69.81
	mtPRS_ML	70.08	70.06	70.10
	UniPRS	64.38	64.38	64.38
	mtSNPPRS_XGB	70.67	70.52	70.81

下载: 导出CSV

参考文献(24)

[1]	CLAUSSNITZER M, CHO J H, COLLINS R, et al. A brief history of human disease genetics[J]. Nature, 2020, 577(7789): 179–189. doi: 10.1038/s41586-019-1879-7.
[2]	MA Ying and ZHOU Xiang. Genetic prediction of complex traits with polygenic scores: A statistical review[J]. Trends in Genetics, 2021, 37(11): 995–1011. doi: 10.1016/j.tig.2021.06.004.
[3]	ZHANG Sai, SHU Hantao, ZHOU Jingtian, et al. Single-cell polygenic risk scores dissect cellular and molecular heterogeneity of complex human diseases[J]. Nature Biotechnology, 2025: 1–17. doi: 10.1038/s41587-025-02725-6.
[4]	LENNON N J, KOTTYAN L C, KACHULIS C, et al. Selection, optimization and validation of ten chronic disease polygenic risk scores for clinical implementation in diverse US populations[J]. Nature Medicine, 2024, 30(2): 480–487. doi: 10.1038/s41591-024-02796-z.
[5]	LOOS R J F. 15 years of genome-wide association studies and no signs of slowing down[J]. Nature Communications, 2020, 11(1): 5900. doi: 10.1038/s41467-020-19653-5.
[6]	ZHU Wensheng and ZHANG Heping. Why do we test multiple traits in genetic association studies?[J]. Journal of the Korean Statistical Society, 2009, 38(1): 1–10. doi: 10.1016/j.jkss.2008.10.006.
[7]	HU Yiming, LU Qiongshi, LIU Wei, et al. Joint modeling of genetically correlated diseases and functional annotations increases accuracy of polygenic risk prediction[J]. PLoS Genetics, 2017, 13(6): e1006836. doi: 10.1371/journal.pgen.1006836.
[8]	GUO Ping, GONG Weiming, LI Yuanming, et al. Pinpointing novel risk loci for Lewy body dementia and the shared genetic etiology with Alzheimer’s disease and Parkinson’s disease: A large-scale multi-trait association analysis[J]. BMC Medicine, 2022, 20(1): 214. doi: 10.1186/s12916-022-02404-2.
[9]	VILHJÁLMSSON B J, YANG Jian, FINUCANE H K, et al. Modeling linkage disequilibrium increases accuracy of polygenic risk scores[J]. American Journal of Human Genetics, 2015, 97(4): 576–592. doi: 10.1016/j.ajhg.2015.09.001.
[10]	GE Tian, CHEN C Y, NI Yang, et al. Polygenic prediction via Bayesian regression and continuous shrinkage priors[J]. Nature Communications, 2019, 10(1): 1776. doi: 10.1038/s41467-019-09718-5.
[11]	KHERA A V, CHAFFIN M, ARAGAM K G, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations[J]. Nature Genetics, 2018, 50(9): 1219–1224. doi: 10.1038/s41588-018-0183-z.
[12]	MAVADDAT N, MICHAILIDOU K, DENNIS J, et al. Polygenic risk scores for prediction of breast cancer and breast cancer subtypes[J]. American Journal of Human Genetics, 2019, 104(1): 21–34. doi: 10.1016/j.ajhg.2018.11.002.
[13]	PRIVÉ F, ARBEL J, and VILHJÁLMSSON B J. LDpred2: Better, faster, stronger[J]. Bioinformatics, 2021, 36(22/23): 5424–5431. doi: 10.1093/bioinformatics/btaa1029.
[14]	ZHONG Peng, ZHANG Chumeng, WU Qinfeng, et al. Shared genetic loci connect cardiovascular disease with blood pressure and lipid traits in East Asian populations[J]. Frontiers in Genetics, 2025, 16: 1635378. doi: 10.3389/fgene.2025.1635378.
[15]	ALLEGRINI A G, SELZAM S, RIMFELD K, et al. Genomic prediction of cognitive traits in childhood and adolescence[J]. Molecular Psychiatry, 2019, 24(6): 819–827. doi: 10.1038/s41380-019-0394-4.
[16]	KRAPOHL E, PATEL H, NEWHOUSE S, et al. Multi-polygenic score approach to trait prediction[J]. Molecular Psychiatry, 2018, 23(5): 1368–1374. doi: 10.1038/mp.2017.163.
[17]	CHUNG W, CHEN Jun, TURMAN C, et al. Efficient cross-trait penalized regression increases prediction accuracy in large cohorts using secondary phenotypes[J]. Nature Communications, 2019, 10(1): 569. doi: 10.1038/s41467-019-08535-0.
[18]	ALBIÑANA C, ZHU Zhihong, SCHORK A J, et al. Multi-PGS enhances polygenic prediction by combining 937 polygenic scores[J]. Nature Communications, 2023, 14(1): 4702. doi: 10.1038/s41467-023-40330-w.
[19]	TRUONG B, HULL L E, RUAN Yunfeng, et al. Integrative polygenic risk score improves the prediction accuracy of complex traits and diseases[J]. Cell Genomics, 2024, 4(4): 100523. doi: 10.1016/j.xgen.2024.100523.
[20]	CHEN T H, CHATTERJEE N, LANDI M T, et al. A penalized regression framework for building polygenic risk models based on summary statistics from genome-wide association studies and incorporating external information[J]. Journal of the American Statistical Association, 2021, 116(533): 133–143. doi: 10.1080/01621459.2020.1764849.
[21]	ZHAI Song, GUO Bin, WU Baolin, et al. Integrating multiple traits for improving polygenic risk prediction in disease and pharmacogenomics GWAS[J]. Briefings in Bioinformatics, 2023, 24(4): bbad181. doi: 10.1093/bib/bbad181.
[22]	王宇翱, 黄叶琪, 李青远, 等. 融合表示学习和知识图谱推理的糖尿病及并发症预测方法[J]. 电子与信息学报. doi: 10.11999/JEIT250798. WANG Yuao, HUANG Yeqi, LI Qingyuan, et al. Integrating representation learning and knowledge graph reasoning for diabetes and complications prediction[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT250798.
[23]	CHEN Tianqi and GUESTRIN C. XGBoost: A scalable tree boosting system[C]. The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, USA, 2016: 785–794. doi: 10.1145/2939672.2939785.
[24]	LAMBERT S A, WINGFIELD B, GIBSON J T, et al. Enhancing the Polygenic Score Catalog with tools for score calculation and ancestry normalization[J]. Nature Genetics, 2024, 56(10): 1989–1994. doi: 10.1038/s41588-024-01937-x.