An important function of flow cytometry is its ability to separate and collect a sub-population of cells, identified by multi-parameter analysis. Classically, this sorting of cells is accomplished as the cells exit from the sample chamber in a liquid jet. Savert showed that when a small jet of fluid was vibrated at the correct frequency the stream could be broken into a series of uniform droplets. In the flow cytometer the sheath stream is broken into a series of uniform droplets by vibrating the sample chamber with a piezoelectric crystal at a high frequency. Cells flowing through the flow cytometer are isolated in these tiny droplets. When the computer detects a cell that satisfies the parameters determined by the operator for sorting, an electrical charge is applied to the droplet. The polarity of the charge-positive or negative, is determined by the sorting criteria. As the charged droplet passes an electrostatic field it is deflected to the right or left, carrying the sorted cell. Extremely pure populations of cells may be sorted at relatively rapid rates.

After identifying an event that satisfies a desired criterion decisions then have to be made when it arrives at the correct position for sorting. With reference to stream in air sorting there is an uncertainty as to exactly which droplet will contain a particular event. Some will occupy a position in the continuous stream that eventually forms a boundary between two droplets and the event could appear in either. Traditionally, three droplets were sorted to overcome this uncertainty. More recently, cytometers with 'phase sorting' are able to predict with more confidence exactly which droplet will contain an event and single droplet sorting is feasible on these instruments. Consider a typical flow rate of 3000 events per second on a stream in air sorter with a drop frequency of 27,000 per second. The average occupancy of a droplet will be 0.09 and, if the events were spaced evenly, 8 empty ones follow each droplet containing a cell. Sorting this idealised distribution would present few problems; a good recovery would be expected of highly purified material. However, not only are the events distributed at random amongst themselves they will also be spaced at random time intervals.

The spacing of events with time will be governed by Poisson distributions. There will be a probability of two events, or more, being placed in the same droplet and of sequential droplets being occupied. These events may either both be of interest or not of interest or, alternatively, there may be one of each. The calculation of these distributions is beyond the scope of this article but it should be obvious that these coincidence effects increase with the flow rate relative to the drop frequency. With a conventional three-droplet sort this would produce a coincidence event with a frequency of 0.16, if these were not sorted (aborted) in the interests of purity the recovery immediately drops to a maximum of 84%. Flow sorters usually give the operator some control over the abort decision. Recovery can be improved, but with the penalty of increased contamination by not aborting coincidence events. The only way to improve recovery and maintain high purity is to slow down the flow rate, decreasing the average drop occupancy and thereby reducing the coincidence frequency and abort rate.

See also FRET sorting

Publications.

JP Biggerstaff, B Weidow, J Dexheimer, G Warnes, J Vidosh, S Patel, M Newman, P Patel. Soluble fibrin inhibits lymphocyte adherence and cytotoxicity against tumor cells:Implications for cancer metastasis and immunotherapy. Clin Appl Thromb Hemost 14(2): 193-202, 2008.

MR Ehrenstein, JG Evans, A Singh, S Moore, G Warnes, DA Isenberg, C Mauri. Compromised function of regulatory T cells in Rheumatoid Arthritis and reversal by anti-TNF therapy. J Exp Med. 200, 277-285, 2004.

P Sabbattini, M Lundgren, A Georgiou, C Chow, G Warnes, N Dillion. Binding of Ikaros to the lambda5 promoter silences transcription through a mechanism that does not require heterochromatin formation. EMBO J. 20(11), 2812-22, 2001.