Table 5 Results of the ABCRF (approximate Bayesian computation with random-forest) model-choice procedure performed to determine which evolutionary process (or combination of processes) best explained the downstream increase in genetic diversity (DIGD) observed for five species with a decreasing polymorphism rate (P95) and/or expected heterozygosity (HS), and increasing local genetic isolation (FST.P), as upstream distance increaseda.

From: Contrasts in riverscape patterns of intraspecific genetic variation in a diverse Neotropical fish community of high conservation value

Species

Number of votes attributed to each process (or combination thereof)

Process(es) selected

RF posterior probability

ASYM

COLON

DIFFNE

ASYM COLON

ASYM DIFFNE

COLON DIFFNE

ASYMCOLONDIFFNE

NULL

Geophagus harreri

6

62

90

11

4

268

55

4

COLONDIFFNE

0.746

Myloplus rhomboidalis

26

117

137

0

0

172

1

47

COLONDIFFNE

0.554

Pseudancistrus barbatus

2

222

67

2

0

196

0

11

COLON

0.595

Serrasalmus eigenmanni

11

53

53

14

3

191

170

5

COLONDIFFNE

0.779

Tometes lebaili

1

17

187

1

1

285

8

3

COLONDIFFNE

0.630

  1. aThe processes examined were downstream-biased migration (ASYM), upstream-directed colonization (COLON), and higher downstream habitat availability (DIFFNE). A combination of these labels indicates a combination of processes. In the NULL model, only the dendritic hydrographic structure influenced the pattern of genetic structure. The out-of-bag error rate in the analysis was of 14.3%. ABCRF analyses were performed by running 500 random-forest (RF) trees in which the eight possible processes received votes. Bold values indicate the process selected.
Back to article page