Broderick Gordon, Fletcher Mary Ann, Gallagher Michael, Barnes Zachary, Vernon Suzanne D, Klimas Nancy G
Department of Medicine, Division of Pulmonary Medicine, University of Alberta, Edmonton, AB, Canada.
Methods Mol Biol. 2012;934:145-64. doi: 10.1007/978-1-62703-071-7_8.
Complex disorders like Gulf War Illness (GWI) often defy diagnosis on the basis of a single biomarker and may only be distinguishable by considering the coexpression of multiple markers measured in response to a challenge. We demonstrate the practical application of such an approach using an example where blood was collected from 26 GWI, 13 healthy control subjects, and 9 unhealthy controls with Chronic Fatigue at three points during a graded exercise challenge. A 3-way multivariate projection model based on 12 markers of endocrine and immune function was constructed using a training set of n = 10 GWI and n = 11 healthy controls. These groups were separated almost completely on the basis of two coexpression patterns. In a separate test set these same features allowed for discrimination of new GWI subjects (n = 16) from unhealthy (n = 9) and healthy control subjects with a sensitivity of 70% and a specificity of 90%.
像海湾战争综合症（GWI）这样的复杂疾病通常难以基于单一生物标志物进行诊断，可能只有通过考虑在应对一项刺激时所测量的多种标志物的共表达情况才能区分。我们通过一个例子展示了这种方法的实际应用，即在分级运动刺激的三个时间点，从26名海湾战争综合症患者、13名健康对照受试者以及9名患有慢性疲劳的非健康对照者身上采集血液。基于12种内分泌和免疫功能标志物构建了一个三向多变量投影模型，使用了n = 10名海湾战争综合症患者和n = 11名健康对照者的训练集。这些组几乎完全基于两种共表达模式被区分开来。在一个单独的测试集中，这些相同的特征能够将新的海湾战争综合症患者（n = 16）与非健康（n = 9）以及健康对照受试者区分开来，灵敏度为70%，特异性为90%。
