sorry, I miscalculated, actually, you should be able to get your
positive
clones if you have around 1-5 in the total cells (means you just need to
sort 5x10^6 cells).
The only potential problem is how you differentiate the positive clones
from
the negative ones. If the sensitivity is good enough, you theoretically
just need one clone. Normally, 5-fold should be safe enough for you to
get
one unless your have a very crappy assay.
It all depends on your assay. I would say 10-fold less of the total
positive clones is a safe number you should be able to get.