Table 1

Log-likelihoods of a linear fit between all ω values and each of 8 common distributions, with likelihood ratio tests for the differences in distributions calculated for the 3 best distributional fits

Distribution

Pearson's r^a

k^b

LRT^c

df

P-value^d

Weibull

.999

2

186.39

2

<10^-6

Gamma

.999

2

201.67

2

<10^-6

Exponential

.998

1

28.29

1

<10^-6

Logistic

.924

2

-^e

-

-

Normal

.923

2

-

-

-

Extreme Value

.903

3

-

-

-

Log-Normal

.854

2

-

-

-

Cauchy

.163

2

-

-

-

a. Pearson's r is the linear correlation of the data to the quartiles, based on the maximum likelihood inferred parameters for each family of distributions.

b. k is the number of free parameters in the distribution.

c. Likelihood ratio test: LRT = 2 * (log-likelihood of metabolic ω values + log-likelihood of non-metabolic ω values) - (log-likelihood for all ω values).

d. Distributed χ².

e. Likelihood ratio test only performed for the best distributional fits.

Hudson and Conant BMC Evolutionary Biology 2011 11:89   doi:10.1186/1471-2148-11-89