In the aqueous two-phase systems (ATPS), the variations in the feedstock characteristics from batch to batch have to be monitored and controlled. In this paper, we have used the distribution analysis of radiolabelled analytes (DARA) as a method to monitor the feedstock variation. The influences of operational variables in yeast processing (e.g. disruption rate, pH, whole cell content and loading) were individually established by the estimation of the tie line length (TLL) by DARA. This information was applied to control and monitor glyceraldehyde 3-phosphate dehydrogenase (G3PDH) during purification in a two-stage ATPS re-extraction system, involving two similar sensitive ATPS. The total protein concentration increased in the top phase as the number of cycles increased, which selectively excluded G3PDH from the top phase to the bottom phase and increased recoveries. This operation was successfully operated on for up to 5-recycles and demonstrated the extended application of DARA in process controls. A fourfold purification of G3PDH was achieved with an average of 85.9% G3PDH recovery. (C) 2011 Elsevier B.V. All rights reserved.