Table 2

Best descriptive rank-abundance form for the viral communities as determined by PHACCS. The error represents the variance weighted sum squared deviation between the experimental and the predicted contig spectra. For each community, the best descriptive function is the one that minimizes the error. The best fit obtained for each rank-abundance form was ranked according to the error in ascending order.

SP

MB

MBSED

FEC

Rank

Model

Error

Rank

Model

Error

Rank

Model

Error

Rank

Model

Error

1

Power law

1.84

1

Power law

2.15

1

Power law

0.0104

1

Power law

9.79

2

Lognormal

1.93

2

Lognormal

2.36

1

Lognormal

0.0104

2

Lognormal

10.2

3

Logarithmic

2.57

3

Logarithmic

2.88

1

Logarithmic

0.0104

3

Logarithmic

10.3

4

Broken stick

10.7

4

Broken stick

14.6

1

Exponential

0.0104

4

Broken stick

52.2

5

Exponential

12.0

5

Exponential

16.2

5

Niche preemption

0.0139

5

Exponential

60.0

5

Niche preemption

12.0

5

Niche preemption

16.2

6

Broken stick

0.0157

5

Niche preemption

60.0

Angly et al. BMC Bioinformatics 2005 6:41   doi:10.1186/1471-2105-6-41