Pattern recognition of monosaccharides via a virtual lectin array constructed by boronate affinity-based pH-featured encoding.

Abstract

Lectin array is an important tool in the fields of carbohydrate chemistry, glycobiology, and glycomics. Because natural lectins are associated with some apparent disadvantages such as tedious purification and easy loss of activity, artificial materials are applied to overcome such shortages by mimicking and replacing lectins in an artificial lectin array, among which boronate affinity-based materials are very outstanding and widely used. However, complicated synthetic works are often involved to design and create boronate affinity-based lectin-mimics. In this work, a facile and novel method was proposed to establish a virtual lectin array based on boronate affinity-based pH-featured encoding for discrimination of monosaccharides by pattern recognition. The dependence of boronate affinity on environmental pH was selected to encode each monosaccharide for feature generation, and the pH-featured encoding was used to construct the virtual lectin array. On the basis of the virtual array, pattern recognition algorithms were applied for data analysis. Monosaccharides were discriminated by principal component analysis, and the relations in the virtual lectin array were unraveled by cluster analysis. In this proof-of-concept work, without complicated synthesis or preparation, the proposed method was successful in mimicking lectin array and discriminating nine elementary monosaccharides found in nature, and it was also a new way of encoding in expanding the applications of boronate affinity-based materials and methods in the field of biomimetics.