This thesis investigates methods for the rapid determination of suitable operating conditions for the primary recovery of antibody fragments from high cell density fermentation broths by two alternative processes: centrifugation and expanded bed adsorption. The methodologies applied involve the use of predictive tools, such as scale-down techniques and simulations, in order to ensure rapid prediction of large-scale process performance. This is followed by the visualisation of suitable processing conditions for recovery of the high cell density cultures investigated using Windows of Operation. Challenges related to protein recovery from high cell density expression systems were identified and addressed. Existing USD clarification approaches lead to an over-prediction of separation performance when tested with high cell density cultures of E. coli whole cells and periplasmically lysed E. coli cells. This was attributed to aggregation effects occurring in the low shear environment of a laboratory centrifuge, which would not be apparent in the settling region of a continuous-flow industrial centrifuge. A modified USD clarification methodology was developed, which resulted in accurate predictions of large-scale performance. This novel USD clarification method was applied to E. coli homogenates and P. pastoris cultures of varying solids concentrations. For these feedstocks, a laboratory-based protocol for the determination of centrifugal dewatering was developed and applied. Windows of Operation were generated, visualising the available operating conditions for a number of industrial centrifuges, when the process was constrained by pre-defined performance and operating criteria. The main challenge identified upon processing of E. coli homogenate by expanded bed adsorption relate to cell-cell and/or cell-adsorbent interactions. These interactions were more prominent in the 1.9 mm ID scale-down column than in the 25 mm ID column, probably as a result of the high particle to column diameter ratio in the scale-down bed. As a consequence, the behaviour of the beds differed in terms of level of expansion, breakthrough profiles, binding capacity and yield. Industrial-scale EBA process performance was investigated using the general rate model to predict the output. This formed the basis for the generation of a series of Windows of Operation, displaying the most suitable combinations of load volume and flow rate for the processing of E. coli homogenates of a range of solids concentrations by EBA. A comparative study was performed based on the identification of suitable operating conditions from the Windows of Operation generated for E. coli homogenate, and suggested that EBA provides higher yields, shorter processing times and greater throughput relative to a more conventional processing route, comprising centrifugation, filtration and packed bed chromatography.