How to calculate the number of transposon mutants for a genomic library?

Jerry M. jerrym at
Tue Aug 16 11:31:31 EST 2005

On Sun, 14 Aug 2005 00:32:12 -0500, "JXB" <jxb at> wrote:

> Hi All,
> I want to make a transposon library in a gram negative bacteria. The
> genome size in ~1.9 Mb and average gene size is 1 kb. Does anyone
> know how I calculate the theoritical number of transposon mutants I
> need to knockout every gene in this bacteria?

For this you have to use some simple statistical formulas derived from
Poisson's Law. Here is a simple one often used to calculate how many
clones are required:

   N = ln(1 - P)/ln(1 - f)

where P is the probability to hit a particular gene, f is the
fractional proportion of the genome represented by a single clone
and N is the number of clones required to get P.

In your case f = 1 / 1900 (you are lucky to have such a small
genome!), so here is a table showing how many (N) clones you will have
to screen to get a particular clone with a probability of P:

    P	  N
  0.100	  200
  0.500	 1317
  0.900	 4374
  0.990	 8748
  0.999	13121

As you see, if you screen about 1300 clones, you have a 50% chance to
get the one particular clone you are interested in, but you will need
to screen about 10 times more to get a 999 in 1000 chance to get it.

Of course, this is in theory. In practice not all the genes have the
same size, there are DNA regions that do not carry any genes, there
are genes that are essential and cannot be disrupted, there are
transposon insertion hot spots, etc. lots of factors that will make
your real statistics certainly not fit the actual theory.

I think the reference for this formula can be found in "The Maniatis"
(aka "The Sambrook"), but I am not sure.

Good luck!


More information about the Methods mailing list