All experiments were performed using the Biomek FX, a multiaxis Liquid Handler platform as the platform for the development of the Reversine magnetic cell separation process. The instrument we used was equipped with two robotic pods. One was a 96-channel pod with a pair of grippers for moving labware around the instrument work area in x, y and z-motions. The second robot arm was a span-8 liquid handler. This pod held a series of eight disposable probes that perform liquid handling operations independent of each other. This arm was used for pipetting, for sample aspirate/dispense operations, for mixing, and when performing blood or reagent transfers. We used 1 ml volume tips when using whole blood. This was the maximum allowable for the instrument. We used 200 mL tips Beckman Coulter, Fullerton, CA) when transferring magnetic beads to blood samples and/or detach-a-bead reagent to bloodcell complexes. Pipetting operations such as aspirating and dispensing of liquids was done at rates of,200�C300 mL/sec. We also added a mechanical mixing option consisting of an Orbital Shaker. This hardware piece was mounted on the instrument deck and attached as an active automated laboratory positioner providing support, power and computer control to allow selectable orbital motions at the 700�C800 rpm rate of the shaker. Similarly, a 90 degree rotating ALP was added to the deck in order to provide position orientation control of the sample vial holder. The ability to adjust the sample holder to the best position for span-8 tip access gave the robot arm the best logistics for optimizing the hardware for high throughput. To control the instrument robotic operations, an x486 PC computer was interfaced to the Biomek FX instrument. The computer operates using the Windows XP Professional OS. A large number of steps are required to automate the magnetic bead-cell separation process, and many of these steps had to supply liquid level position information supplied as feedback to the software Method in order to keep red blood cell contamination to a minimum during washing steps. We learned through trial and error that the Adriamycin capacitive detection liquid level sensing technology supplied with the Biomek FX instrument was unable to provide this accuracy on a consistent basis when doing whole blood samples. This made it necessary for us to write our own computer software program that could address this problem.