Table 1

Regression analysis. Results of step-wise multiple linear regression based on the AIC criterion. Left side: c-score in dependence on all qualitative variables: equation (1), (2), (3), or (4); equation (5) or (6); average, minimum, or maximum of asymmetric distance values; BLASTN, TBLASTX, BLASTN after TBLASTX, or matrix combination; UPGMA, NJ, BIONJ, FastME, or STC. Right side: δ value in dependence on all variables. Only those explanatory variables judged as significant in at least one of the two approaches are included; those not shown were removed in the course of the step-wise procedures and did not contribute to the final regression model.

Regression analysis

cscore (adjusted R² = 0.775)

δ value (adjusted R² = 0.888)

explanatory var.

coefficient

standard error

t value

P(x > |t|)

coefficient

standard error

t value

P(x > |t|)

(Intercept)

0.032509

0.010895

2.984

0.00292

0.533266

0.008668

61.522

< 2^-16

Plastids

0.123627

0.006290

19.655

< 2^-16

-0.172317

0.005779

-29.820

< 2^-16

BLASTN + TBLASTX

0.047477

0.008895

5.337

1.18^-07

-0.044311

0.008172

-5.422

1.84^-07

Matrix Averaging

0.024675

0.008895

2.774

0.00565

not significant

TBLASTX

0.044962

0.008895

5.055

5.18^-07

-0.041812

0.008172

-5.116

7.82^-07

Eq. (6)

not significant

0.043077

0.005779

7.455

3.48^-12

Eq. (2)

0.437622

0.008895

49.197

< 2^-16

-0.184422

0.008172

-22.567

< 2^-16

Eq. (3)

0.380340

0.008895

42.757

< 2^-16

-0.125118

0.008172

-15.310

< 2^-16

Eq. (4)

0.247183

0.008895

27.788

< 2^-16

-0.088696

0.008172

-10.853

< 2^-16

STC

-0.040611

0.009945

-4.083

4.81^-05

UPGMA

-0.029744

0.009945

-2.991

0.00285

Auch et al. BMC Bioinformatics 2006 7:350   doi:10.1186/1471-2105-7-350