Proximity ligation is a powerful technique to measure minute concentrations of target protein with high specificity, and it has been demonstrated to be effective on a wide variety of protein targets. The proximity ligation assay (PLA) technique is shown to be compromised by the amplification of a nonspecific fluorescent product that is not indicative of protein presence, which was previously unidentified in a published procedure. This result illuminates the complexity of designing the optimal PLA and the possibility of using a size-based separation to increase the reliability of PLAs in general. Nucleic acid controls were developed to optimize the assay, which led to a novel endpoint detection method that exploits microchip electrophoresis to size the products. This method provides a greater ability to distinguish a between the target protein's signal and noise in a PLA. The utility of the PLA is demonstrated by the detection of human pathogenic Escherichia coil O157:H7 bacteria, a pathogen at the root of many recent life-threatening food poisoning outbreaks. The results of the PLA show a detection limit of 100 E. coil O157:H7 cells with minimal cross-reactivity with gram positive control Staphylococcus aureus bacteria. The advantages of miniaturizing this process are the 100-fold reduction in volume, greatly reducing reagent requirements, and doubling of the thermocycling speed via noncontact infrared heating. This work, consequently, adds to the understanding of background fluorescence in PLAs, provides a method for evaluating nonspecific amplification, and shows that a qualitative PCR response indicative of the presence protein can be achieved with PLA.