. 2005 Jul 15;33(12):3907–3918. doi: 10.1093/nar/gki699

Table 2.

Ratios of occurrences of the A- and G-tracts of different length in the E.coli genome and in random sequences with the same dinucleotide content^a

Tract length (bp)	A-tracts^b		G-tracts^b
	A_nT_m	A_n:T_n	G_nC_m	G_n:C_n
4	0.88^***	1.05^***	0.83^***	0.71^***
5	1.10^***	1.34^***	0.64^***	0.55^***
6	1.14^***	1.48^***	0.44^***	0.34^***
7	1.09^**	1.26^***	0.32^***	0.28^***
8	1.14^***	0.95	0.15^***	0.20^***
9	0.78^***	0.22^***	0.05^***	0.09
10	0.28^***	0.00	0.12^**	0.46

^aThe ratios ‘genome/random’ are averaged over 10 implementations of a random sequence with the same dinucleotide composition as in the genome. **Significance level P < 0.01 and ***significance level P < 0.001 (t-test).

^bThe results for tracts in the form A_nT_m or G_nC_m are shown separately from the tracts containing purines in one strand and pyrimidines in the other strand (A_n:T_n or G_n:C_n).