Skip to main content

Table 6 Summary of results for the chicken data set, representing 20 breeds

From: Characterization of a Bayesian genetic clustering algorithm based on a Dirichlet process prior and comparison among Bayesian clustering methods

Program

λ

Number of clusters

Differences from the partition that was determined from breeds

HWLER

23

Breed 21 was divided into two clusters (14 and 16 individuals), breed 121 was divided into four clusters (1, 1, 3, and 25 individuals), and breeds 44 and 45 shared a cluster.

DPART

Inferred (unique)

23

Same as HWLER

 

Inferred (single)

22

Breed 121 was divided into four clusters (1, 1, 3, and 25 individuals). Breeds 44 and 45 shared a cluster.

 

0.05

23

Same as HWLER

 

0.5

20

Breed 121 was divided into two clusters (5 and 25 individuals), breeds 44 and 45 shared a cluster, an individual in breed 5 shared a cluster with breed 50, and an individual in breed 16 shared a cluster with breed 5.

 

1

17

Breeds 5 and 6, 18 and 37, and 44 and 45 shared different clusters respectively. Three individuals in breed 102 shared a cluster with breed 33. An individual in breed 5 shared a cluster with breed 50.

 

3

9

Breeds 5, 16, 18, 21, 37 and 3402 shared a cluster. Breeds 33, 44, 45, 51 and an individual in breed 102 shared a cluster. Breed 13, 26, 42, 50, and an individual each in breeds 5 and 102 shared a cluster.