Ge-SAND: an explainable deep learning-driven framework for disease risk prediction by uncovering complex genetic interactions in parallel

Ye, Lihang; Zhang, Liubin; Tang, Bin; Liang, Junhao; Tan, Ruijie; Jiang, Hui; Peng, Wenjie; Lin, Nan; Li, Kun; Xue, Chao; Li, Miaoxin

doi:10.1186/s12864-025-11588-9

Table 1 Performance of different models for the simulation (QD, N = 2,000), Crohn’s disease, Schizophrenia, and Alzheimer’s disease testing sets

From: Ge-SAND: an explainable deep learning-driven framework for disease risk prediction by uncovering complex genetic interactions in parallel

Dataset	Model	AUC -ROC	AUC-PR	KS	MCC	ACC	Recall	Precision	F1 score
Simulation (N = 2,000 QD)	Ge-SAND	0.673	0.678	0.310	0.319	0.655	0.770	0.626	0.691
	SVM	0.579	0.552	0.170	0.170	0.585	0.700	0.569	0.628
	Naive Bayes	0.585	0.578	0.160	0.160	0.580	0.590	0.578	0.584
	Lasso	0.622	0.603	0.210	0.225	0.600	0.830	0.569	0.675
	Ridge	0.586	0.574	0.180	0.171	0.585	0.530	0.596	0.561
	Random Forest	0.514	0.530	0.080	0.090	0.540	0.310	0.574	0.403
	XGBoost	0.594	0.597	0.220	0.211	0.605	0.550	0.618	0.582
	MLP	0.626	0.655	0.240	0.277	0.620	0.370	0.740	0.493
	CNN	0.446	0.487	0.200	0.085	0.525	0.930	0.514	0.662
	LSTM	0.516	0.534	0.090	0.116	0.545	0.230	0.622	0.336
Crohn's Disease	Ge-SAND	0.700	0.637	0.386	0.393	0.684	0.860	0.636	0.731
	SVM	0.541	0.555	0.140	0.131	0.561	0.737	0.546	0.627
	Naive Bayes	0.503	0.506	0.105	0.093	0.544	0.386	0.564	0.458
	Lasso	0.508	0.512	0.123	0.109	0.553	0.421	0.571	0.485
	Ridge	0.551	0.576	0.158	0.193	0.570	0.228	0.722	0.347
	Random Forest	0.585	0.571	0.193	0.199	0.597	0.474	0.628	0.540
	XGBoost	0.496	0.515	0.105	0.113	0.553	0.368	0.583	0.452
	MLP	0.553	0.564	0.158	0.158	0.579	0.579	0.579	0.579
	CNN	0.576	0.593	0.158	0.171	0.579	0.772	0.557	0.647
	LSTM	0.524	0.518	0.105	0.117	0.553	0.772	0.537	0.633
Schizophrenia	Ge-SAND	0.646	0.637	0.295	0.282	0.641	0.641	0.641	0.641
	SVM	0.526	0.574	0.115	0.185	0.558	0.167	0.765	0.274
	Naive Bayes	0.583	0.602	0.231	0.238	0.609	0.410	0.681	0.512
	Lasso	0.504	0.546	0.090	0.079	0.539	0.423	0.550	0.478
	Ridge	0.559	0.593	0.141	0.131	0.564	0.462	0.581	0.514
	Random Forest	0.554	0.585	0.154	0.269	0.577	0.167	0.929	0.283
	XGBoost	0.510	0.542	0.077	0.100	0.539	0.859	0.523	0.651
	MLP	0.571	0.592	0.205	0.228	0.603	0.821	0.571	0.674
	CNN	0.526	0.537	0.115	0.119	0.558	0.679	0.546	0.606
	LSTM	0.520	0.536	0.128	0.159	0.564	0.269	0.656	0.382
Alzheimer's Disease	Ge-SAND	0.716	0.717	0.380	0.381	0.690	0.709	0.683	0.695
	SVM	0.561	0.575	0.103	0.122	0.552	0.817	0.534	0.646
	Naive Bayes	0.542	0.569	0.099	0.173	0.547	0.127	0.794	0.219
	Lasso	0.525	0.559	0.070	0.110	0.535	0.150	0.653	0.244
	Ridge	0.561	0.577	0.132	0.127	0.535	0.150	0.653	0.244
	Random Forest	0.593	0.596	0.155	0.155	0.578	0.545	0.583	0.563
	XGBoost	0.658	0.663	0.253	0.249	0.624	0.624	0.624	0.624
	MLP	0.573	0.586	0.131	0.150	0.566	0.324	0.627	0.427
	CNN	0.464	0.475	0.099	0.042	0.512	0.925	0.506	0.654
	LSTM	0.514	0.511	0.047	0.047	0.523	0.559	0.522	0.540

Back to article page

ISSN: 1471-2164

Contact us

Submission enquiries: bmcgenomics@biomedcentral.com
General enquiries: ORSupport@springernature.com

BMC Genomics

Contact us