Skip to main content
. 2017 Jan 25;18:106. doi: 10.1186/s12864-017-3486-z

Table 4.

Selection patterns in recombination genes in maize detected using the likelihood ratio test

Gene Selection modelsa % of codons
M0 vs. M3
(2Δl) b
M1 vs. M2
(2Δl) c
M7 vs. M8
(2Δl) d
Under purifying selection Under positive selection Evolving neutrally
Dmc1 15.32** 4.12 6.14 100 - -
Mlh1 50.48** 19.46** 20.54** 93.3 6.7 -
Mre11A 1.06 0.46 0.12 100.0 - -
Mre11B 24.64** 26.46** 14.23** 96.5 3.5 -
Msh4 0.00 0.00 0.00 100.0 - -
Mus81-1 22.64** 15.62** 15.74** 97.3 2.7 -
Rad51A1 0.00 0.00 0.00 100.0 - -
Rad51A2 43.76** 18.94** 20.62** 98.7 1.3 -
Recq4 0.00 0.00 0.00 100.0 - -
Spo11-1 0.00 0.00 0.00 100.0 - -
Spo11-2 80.52** 33.78** 35.18** 98.2 1.8 -

a Likelihood ratio tests of selection models. Ratios statistically significant at P = 0.01 are denoted with **

b Model M0 (one ratio) assumes a single dN/dS ratio across all sites in the gene’s coding region. Model M3 (discrete) assumes discrete classes of sites with different ω values. df = 4

c Model M1 (neutral model) assumes all sites are either under purifying selection (ω < 0) or evolving neutrally (ω = 1). Model M2 adds a third category of sites under positive selection (ω > 1). df = 2

d Model M7 (beta model) assumes that ω ranges from 0 (strong negative selection) to 1 (neutral evolution) and varies among sites in the gene’s coding region according to the beta distribution. Model M8 (beta& ω) similarly to M7 assumes that ω varies among sites but allows that, in addition to ranging from 0 to 1, ω may take values > 1 (positive selection). df = 2