net.maizegenetics.analysis.clustering

Class HaplotypeClusterer

net.maizegenetics.analysis.clustering.HaplotypeClusterer

public class HaplotypeClusterer

HaplotypeClusterer clusters haplotypes by distance as defined in the Haplotype class. This class can be extended to apply different clustering rules.

Constructor Detail

HaplotypeClusterer

public HaplotypeClusterer(java.util.List<kotlin.Array[]> haplotypes)

Parameters:

haplotypes - a List of haplotype or genotype sequence to be clustered

HaplotypeClusterer

public HaplotypeClusterer(java.util.ArrayList<net.maizegenetics.analysis.clustering.Haplotype> haplotypes)

Parameters:

haplotypes - a List of Haplotypes to be clustered

Method Detail

makeClusters

public void makeClusters()

Groups a list of Haplotypes into clusters. Clusters are created so that all Haplotypes in a cluster have 0 pairwise distance. Because of missing data a Haplotype can be assigned to more than one cluster.

removeFirstHaplotypes

public HaplotypeCluster removeFirstHaplotypes(int maxdistance)

Removes all haplotypes within maxdistance of the haplotype of the first cluster. After the haplotypes have been removed clusters are remade and sorted.

Parameters:

maxdistance - all haplotypes within maxdistance or less of the haplotype of the first cluster are removed from the haplotypeList

sortClusters

public void sortClusters()

Sorts clusters according to HaplotypeCluster sort order.

getNumberOfClusters

public int getNumberOfClusters()

Returns:

org.jetbrains.dokka.NodeRenderContent@7281c174: the number of clusters

getClusterSizes

public kotlin.Array[] getClusterSizes()

Returns:

org.jetbrains.dokka.NodeRenderContent@7a573bb1: the number of Haplotypes contained in each cluster

getClusterScores

public kotlin.Array[] getClusterScores()

After the initial cluster formation a Haplotype score equals the 1 / (number of clusters to which it belongs). Merging does not update the cluster score.

Returns:

org.jetbrains.dokka.NodeRenderContent@6ecbccf4: the score for each cluster.

getClusterList

public java.util.ArrayList<net.maizegenetics.analysis.clustering.HaplotypeCluster> getClusterList()

Returns:

org.jetbrains.dokka.NodeRenderContent@4b92805e: List of clusters

clusterDistanceDistinctHaplotypes

public static int clusterDistanceDistinctHaplotypes(HaplotypeCluster cluster0,
                                                    HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@37647239: the minimum of the number of haplotypes in cluster0 not in cluster1 and the number of haplotypes in cluster 1 not in cluster0

clusterDistanceDistinctHaplotypeProportion

public static double clusterDistanceDistinctHaplotypeProportion(HaplotypeCluster cluster0,
                                                                HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@74ad5dd7: the minimum of the number of haplotypes in cluster0 not in cluster1 and the number of haplotypes in cluster 1 not in cluster0 divided my the minimum number of haplotypes in the two clusters

clusterDistanceClusterHaplotypeDiff

public static int clusterDistanceClusterHaplotypeDiff(HaplotypeCluster cluster0,
                                                      HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@443183da: the number of alleles difference between the cluster haplotypes

clusterDistanceClusterDiffProportion

public static double clusterDistanceClusterDiffProportion(HaplotypeCluster cluster0,
                                                          HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@1bfe7b72: the number of alleles difference between the cluster haplotypes divided by the number of nonmissing sites

clusterDistanceMaxPairDiff

public static int clusterDistanceMaxPairDiff(HaplotypeCluster cluster0,
                                             HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@f9f278e: the maximum of the pairwise difference between individual haplotypes in the two clusters

clusterDistanceAveragePairDiff

public static double clusterDistanceAveragePairDiff(HaplotypeCluster cluster0,
                                                    HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@1142804f: the average of the pairwise differences between the haplotypes that are not shared between clusters

clusterDistanceTotalPairDiff

public static int clusterDistanceTotalPairDiff(HaplotypeCluster cluster0,
                                               HaplotypeCluster cluster1)

Parameters:

cluster0 -

cluster1 -

Returns:

org.jetbrains.dokka.NodeRenderContent@5925eeee: the total of the pairwise differences between the haplotypes that are not shared between clusters

getMergedClusters

public static java.util.ArrayList<net.maizegenetics.analysis.clustering.HaplotypeCluster> getMergedClusters(java.util.ArrayList<net.maizegenetics.analysis.clustering.HaplotypeCluster> candidateClusters,
                                                                                                            int maxdiff)

Merges clusters whose maximum pairwise difference is less than maxdiff. Clusters are tested sequentially. That is, if two clusters are merged, they become the new head cluster against which remaining clusters are tested for merging.

Parameters:

candidateClusters - an ArrayList of HaplotypeClusters

maxdiff -

Returns:

org.jetbrains.dokka.NodeRenderContent@31683b53: merges clusters whose maximum pairwise difference is less or equal to maxdiff

mergeClusters

public void mergeClusters(int maxdiff)

Merges clusters whose maximum pairwise difference is less than maxdiff. Clusters are tested sequentially. That is, if two clusters are merged, they become the new head cluster against which remaining clusters are tested for merging.

Parameters:

maxdiff -

doMerge

public static boolean doMerge(HaplotypeCluster c0,
                              HaplotypeCluster c1,
                              int maxdiff)

Tests whether two clusters are less than or equal to maxdiff distant. Uses clusterDistanceMaxPairDiff() to calculate distance.

Parameters:

c0 - a cluster

c1 - another cluster

maxdiff -

Returns:

org.jetbrains.dokka.NodeRenderContent@473dd9a7: true if clusters are maxdiff or less distant, false otherwise

mergeTwoClusters

public static void mergeTwoClusters(HaplotypeCluster c0,
                                    HaplotypeCluster c1)

Merges two clusters, c0 and c1.

Parameters:

c0 -

c1 -

removeClusterHaplotypesFromOtherClusters

public void removeClusterHaplotypesFromOtherClusters(int clusterIndex)

For this cluster, remove all of its haplotypes from other clusters, adjust the cluster scores, and re-sort the clusters. As a result, cluster size will equal cluster score for this cluster.

Parameters:

clusterIndex - the index of a cluster in the cluster list

moveAllPossibleHaplotypesToCluster

public void moveAllPossibleHaplotypesToCluster(int clusterIndex,
                                               boolean fromClustersWithHigherIndexOnly,
                                               int maxdiff)

For the indexed cluster, move all haplotypes consistent with the cluster haplotype to this cluster from any other cluster

Parameters:

clusterIndex - the index of a cluster in the cluster list

moveAllHaplotypesToBiggestCluster

public void moveAllHaplotypesToBiggestCluster(int maxdiff)

For each cluster, from largest to smallest, move all consistent haplotypes from other clusters to that cluster

removeHeterozygousClusters

public void removeHeterozygousClusters(int maxHetSites)

recalculateScores

public void recalculateScores()

Recalculates the scores of the clusters in the cluster list. Removes any clusters with a score of 0.