Skip to main content

Table 12 Correlation analyses for 1000-taxon model conditions

From: MRL and SuperFine+MRL: new supertree methods

 

Scaffold Density

Statistic

20

50

75

100

MRP Score

0.770

0.908

0.968

0.991

MRL Score

0.871

0.935

0.976

0.988

Sum-FN

0.762

0.907

0.966

0.992

  1. We show the average Spearman's rank correlation coefficient between different statistics and the FN error rates of trees generated around each of the six estimated supertrees for the 1000-taxon model conditions. For each of the estimated six supertrees, 100 trees were generated using a p-ECR move, for a total of 606 trees (600 p-ECR trees plus 6 supertrees) per replicate. MRP score is the MP score of the estimated tree with respect to the MRP matrix. MRL score is the negative log-likelihood score of the estimated tree with respect to the MRP matrix. Sum-FN is the sum of the bipartitions in the source trees not present in the estimated tree divided by the total number bipartitions in the source trees. Coefficients with larger magnitude represent stronger correlation between the test statistic and FN error rates. The largest correlation coefficient for each scaffold density is shown in bold.