2018 Volume 12 Issue 4 Pages 360-368
It was widely accepted that stroke onset was the result of interactions between environment and genetic factors. However, the combined biomarkers covering environment and genetic factors and their interplay information in stroke were still lacking. In this study, we proposed a framework to identify the targeting or indicating role each factor played in the combined stroke biomarkers. A combined set of 36 biomarkers were identified based on evaluation and importance scores. Validations on three independent microarray data sets justified that the obtained markers were pervasively effective in discriminating stroke patients of different stages from healthy people on genetic levels. 8 and 3 genetic factors were identified as biomarkers in the acute and recovery phases of stroke, respectively. For example, the expression changing of SERPINH1 only appeared in the acute phase of stroke showing its targeting role in the combined biomarker. Compared with this, 11 genetic factors such as MMP9 were found to be differentially expressed in both acute and recovery phases of stroke showing their indicating roles in stroke. Functional analyses further revealed that the biomarkers could be grouped into 4 closely related processes of stroke including prevention, occurrence, processing, and recovery, respectively. These results indicated that the adoption of interactions between environment and genetic factors would be helpful in selecting robust and biologically relevant biomarkers, which cast a new insight for stroke biomarker identification.