Python Programming for Biology: Bioinformatics and Beyond

Figure 21.2.  Probabilities partitioning a unit line. Probability values for the different

possible outcomes can be imagined to partition a region of length one. Here we illustrate

this for the probabilities for the colour of mouse offspring resulting from a genetic cross

(between a pure white strain and a pure black strain) where black is three times more

likely than white.

For a biological system we may have a good model of how we expect things to behave,

i.e. that we know what the probabilities are. For the genetics example illustrated in

Figure

21.2

we might assume that the probability is ¾ for having black offspring and ¼ for white

offspring,  and  likewise  for  a  position  in  DNA  we  might  say  that  the  probability  of  each

base is ¼ and thus for a pair of bases is

1

/

16

.

We should not forget, however, that these values stem from idealised models, so that in

reality  the  actual  probabilities  are  not  neat  whole-number  fractions.  Accordingly,  for  a

DNA nucleotide position it is only an approximation to say that each base is equally likely.

In reality in any given genome there will be more G and C bases than A and T bases, or

vice  versa.  It  should  be  noted  that  if  we  consider  both  DNA  strands  then  because  of  the

base-pairing rules the numbers of G and C will be the same, and thus also the numbers of

A and T. Naturally this means that the probability of finding a base at a random position is

also equal within these pairs.