This distance is the number of sites at which the two compared sequences differ. If you are using the pairwise deletion option for handling gaps and missing data, it is important to realize that this count does not normalize the number of differences based on the number of valid sites compared, if the sequences contain alignment gaps . Therefore, we recommend that if you use this distance you use the complete-deletion Option.
For this distance, MEGA provides facilities for computing the following quantities:
d: Transitions + Transversions: Number of different nucleotide sites.
s: Transitions only: Number of nucleotide sites with transitional differences.
v: Transversions only: Number of nucleotide sites with transversional differences.
R = s/v: Transition /transversions ratio.
L: No of valid common sites: Number of compared sites.
Formulas for computing these quantities and their variances are as follows.
Var(d) = nd(L - nd)/L
Var(s) = s(L - s)/L
Var(v) = v(L - v)/L
R = s/v
Var(R) = [c12P + c22Q – (c1P + c2Q)2)]/L
where c1 = 1/s and c2 = -s/v2
P and Q are the proportion of sites showing transitional and transversional differences, respectively.
See also Nei and Kumar (2000), page 33.