Bioprocessing case study : application of fractional factorial and taguchi parameter design for optimisation of multi-phase in-vitro erythropoiesis process

Abstract

In humans, studies of multi-phase in-vitro erythropoiesis are hampered by difficulties in obtaining sufficient and robust numbers of erythroid progenitors and mature erythrocytes. In this study, classical fractional factorial and Taguchi parameter design were performed to attain a systematic understanding of the significant factors. A resolution IV fractional factorial design used for initial screening with seven factors (Cytokines) viz., SCF, TPO, FL, IL-3, IL6, GM-CSF and EPO tested. Using ANOVA, Normal plot and Half-normal plot and fold expanded as response, the significant factors (SCF and EPO) were identified and combined with IGFII for as control factors and umbilical cord blood donor to donor variation as noise factor. Next, Taguchi parameter design was implemented over 3 Phases (Day 0-6, 6-10 and 10-16) for more in depth study and factor optimization. A L8 orthogonal array with 3 factors of control at 2 levels each and 4 interactions was used. With ANOVA, Half normal plot and Response table, IGFII, EPO, IGFII x EPO ; SCF and EPO; SCF and EPO were identified as the significant factors/interactions at Phases I, II and III respectively. From the S/N Plot, the optimum parameter levels were SCF 10, 110, 110ng/ml IGFII 100, 0,0ng/ml and EPO 5, 5,5 IU/ml at phases I,II and III respectively. A systematic framework for use of fractional factorial and Taguchi parameter design for a multi-phase in-vitro erythropoiesis is shown to be feasible and may serve as framework for future efficient studies on complex biological processes.