One 1/2 liter or so bag of human serum is obtained by either separating and then pooling the serum from several whole-blood donations, or from a single individual by way of a long and costly process of phoresis (I used to undergo such a procedure when I donated blood).

If you had to give a blood test to 3000
people to test for the presence of
Hepatitis C, how would you do it?
Would you test each person individually,
which would require 3000 tests?
That could be expensive, especially if
each test were to cost $100. How can
you use less tests? One way is to pool
the samples. What sample size would
you use? If you pooled the blood of
100 people and the test came out negative,
then one test was used instead
of 100. But if the test came out positive,
then you would need to test each
one, which now requires 1  100 
101 tests. Maybe the sample size to
use is 200. Could it be 300? And the
likelihood of a positive test must play
a role as well. Sounds like a very hard
problem.



--------------------------------------------------
Note added at 8 hrs (2010-11-07 18:57:24 GMT)
--------------------------------------------------


Probe colective ?