Background We studied the consistency between two classification systems for categorizing patients: traditional Chinese medicine (TCM) constitution-based methods, versus?genetic clustering. Genetic classification in constitutional identification was also evaluated. Methods A TCM physician evaluated the constitution of each patient, according to four examinations (inspection, auscultation-olfaction, interrogation, and palpation). Those who met the criteria for Yang-deficient, Yin-deficient, and balanced constitutions were enrolled in the study. Peripheral blood samples were obtained from the participants, and peripheral blood mononuclear cells were separated from the samples within 2 hours. Total RNA extraction from the white blood cells was performed; and an Affymetrix HG-U133 Plus2.0 array was used to determine the peripheral blood gene expression profiles. The samples were classified using a support vector machine genetic classifier, and the “leave-one-out” method was used for validation. Results The global gene expression profiles of 32 samples were grouped into three categories, and the samples in each of the gene categories corresponded with the three constitution categories. The three constitution types were distinguished using the genetic classifier with 165 genes. The accuracy of the prediction classification was greater than 95% using mathematical method. Conclusions Participants with Yin-deficient, Yang-deficient, and balanced constitutions have varying physical characteristics and gene expression patterns. Additionally, the results from TCM constitution classification matched those obtained by genetic classification. Finally, our preliminary gene classifier distinguishes among Yin-deficient, Yang-deficient, and balanced constitutions, and provides a methodological basis for identifying the different constitutions.
展开▼
