Protein A chromatography is widely employed for the capture and purification of antibodies and Fc-fusion proteins. Due to the high cost of protein A resins, there is a significant economic driving force for using these chromatographic materials for a large number of cycles. The maintenance of column performance over the resin lifetime is also a significant concern in large-scale manufacturing. In this work, several statistical methods are employed to develop a novel principal component analysis (PCA)-based tool for predicting protein A chromatographic column performance over time. A method is developed to carry out detection of column integrity failures before their occurrence without the need for a separate integrity test. In addition, analysis of various transitions in the chromatograms was also employed to develop PCA-based models to predict both subtle and general trends in real-time protein A column yield decay. The developed approach has significant potential for facilitating timely and improved decisions in large-scale chromatographic operations in line with the process analytical technology (PAT) guidance from the Food and Drug Administration (FDA). Biotechnol. Bioeng. 2011;108: 59-68. (C) 2010 Wiley Periodicals, Inc.