Similarity assessment by multivariate statistics method based on the distance between biosimilar and originator

The development of biosimilar products or follow-on biologics has been flourishing in recent years because of their lower price than the originators. In this study, a multivariate data analysis method based on JMP software was proposed to assess the glycosylation pattern similarity of antibody candidates from different conditions in optimization experiments with a reference. A specific distance was generated by this method and indicated the glycoform similarity between the biosimilar and the reference. This method can be applied to analyze the similarity of other physicochemical and functional characteristics between follow-on biologics and originators. Then, the design of experimental methods can be realized to optimize the conditions of cell culture to attain similar antibody candidates. A higher concentration of GlcNAc added to the basal media made the glycan of the antibody more similar to the glycan of the reference in this study.