This paper is devoted to discussing two research patterns of biological basis of traditional Chinese medicine syndrome and presenting a research strategy for data mining methods. It points out that data mining methods which are based on feature selection are better fit for investigating biological basis of traditional Chinese medicine syndrome. Based on such a discussion, the concept of "characteristic pattern" is proposed to bridge the gap between "golden index" and biological basis of traditional Chinese medicine syndrome. This paper presents a novel research avenue for investigating biological basis of traditional Chinese medicine syndrome.