缺血性脑卒中是我国居民的重要致死和致残原因。如能准确预测缺血性脑卒中患者的不良结局发生风险，将对该病的个体化防治和预后改善具有重要意义。本项目利用新构建的统计分析模型，通过有效识别代谢组学高维数据中的特定标志物组群，提高疾病预后预测的准确度。主要研究内容包括：缺血性脑卒中代谢组学检测、自适应进化的免疫遗传算法构建、模型参数优化与性能评价、缺血性脑卒中预后代谢标志物组群识别。本研究重点给出一种新的自适应免疫遗传算法（SA-IGA），该方法采取了多种自适应进化策略，不仅能够高效搜索医学高维空间中的目标标志物组群，而且可以给出多样化的标志物组群，并能有效利用医学先验信息。同时，通过结合实际的缺血性脑卒中代谢组学研究，期望给出具有高预测能力和实际应用价值的预后代谢标志物组群，用于缺血性脑卒中患者的不良结局预测。

Ischemic stroke is an important cause of death and disability of Chinese citizens. Predicting the adverse outcomes of ischemic stroke patients accurately is of great significance for the individualized prevention and therapy and prognosis improvement of this disease. This study will make use of the newly constructed statistical analysis model and effectively identify specific biomarker panels in high-dimensional metabolomic data, so as to promote disease prognosis prediction accuracy. The main research contents include: metabolomics detecting for ischemic stroke, construction of the self-adaptively evolved immune genetic algorithm, model parameter optimization and performance assessment, identification of prognostic metabolic biomarker panels for ischemic stroke. This study lays stress on providing a new self-adaptive immune genetic algorithm (SA-IGA). Through adopting several self-adaptive evolving strategies, this method can search the targeted biomarker panels from medical high-dimensional space in a highly effective manner, maintain the diversity of biomarker panels, and effectively make use of medical priors. Meanwhile, it is expected to provide the prognostic metabolic biomarker panels with high predictive ability and practical value, which can be used for predicting the adverse outcomes of ischemic stroke.