Cell Sorting Purity

Fri Jan 21 07:50:45 EST 1994

I've done cell sorting for about 14 years with a Coulter Epics V instrument.  I
have always sorted using a "three droplet" setup.  In my experience, given
the sorter is set up correctly (phase and delay settings) and if there is not a
major static electricity interference problem (that might cause sorted stream
deflections), there are two factors that I feel are very important in getting
the best purity. 
  1- Separation of positives from negatives in the sorting regions.  That is, 
the farther apart the two populations appear in the histogram or bit map 
regions, the higher the purity of the resulting sorted populations.  This I 
believe is mainly due to Gaussian overlaps of populations close in intensities.
You did not say how close the regions are, but If they are close, I think that 
it really adds much to your woes.  If you could increase the difference in the
signal between the positives and negatives, without adding noise, I'd predict
your purity will go up markedly. If you are sorting fluorescent populations,
consider going to a brighter label for your positives (PE). 
  2- Sort rate.  My experience is that you can not get close to the 
instrument manufacturers' suggested maximum sort rates without a dramatic 
decrease in purity.  There is a real trade-off in sort rates to purity in the 
older instruments.  Jim Wood of Coulter published a very detailed evaluation of 
this effect.  You might consider getting in touch with him for the details.  My 
own experience is that I try never to exceed sort rates greater than 3,000
cells (input - not actual cell deposition rate) per second.  When sorting a
minor population, like the one you are dealing with, I would consider slowing
it down even below this rate.  Of course, have your "coincidence 
detection/abort" switched on.

Unfortunately, it seems that you can't have it both ways.  That is
you can't have high purity AND an excellent yield of cells per unit time when
you are dealing with a DIM, low percentage starting material.  You may have to
sacrifice in the number of resulting cells for achieving higher purities.  Good

Phil Marder
Lilly Research Labs
Indianapolis IN 46285

From: MARDER PHILIP                 (MCVAX0::MARDER)

To:   FOREIGN TRANSPORT ADDRESSEE   (MCDEV1::IN%"gpingul at fred.fhcrc.org")
cc:   FOREIGN TRANSPORT ADDRESSEE   (MCDEV1::IN%"kelley at flowcyt.cyto.purdue.edu")
      MARDER PHILIP                 (MCVAX0::MARDER)

More information about the Cytometry mailing list