inspect_window_dbScan-hbr-method: Clustering with the dbscan algorithm
In haploReconstruct: Reconstruction of Haplotype-Blocks from Time Series Data

Description Usage Arguments Details Author(s) See Also

Performs clustering with the dbscan (Density-based spatial clustering of applications with noise) for all SNPs in a window after filtering (i.e. initialize_SNP_time_series) and visualizes found clusters based on principal components. NOTE: The clusters identified here are not nessesarily identical with the clusters identified with average linkage clustering. Therefore if haplotype reconstruction was done using average linkage clustering the clusters shown here can be different from clusters of identified haplotype blocks.

1
2
3

## S4 method for signature 'hbr'
inspect_window_dbScan(object, window,
  minPts = object@min.cl.size, eps, plotkNN = T, plotCluster = T)

`object`	object of class `hbr` containing the results of reconstructed haplotype-blocks
`window`	Number of the window, which should be inspected
`minPts`	Number of minimum points in the eps region (for core points). (default `min.cl.size` used for haplotype block reconstruction)
`eps`	The size of the epsilon neighborhood. Defines the distance between samples that built up a core cluster, for details see ?`dbscan` of the `dbscan` package. (This value is only used for the `plotCluster` functionality.)
`plotkNN`	Boolean indicating if the kNN distance plot should be printed , for details see ?`kNNdistplot` of the `dbscan` package. Briefly, the y-value where a bent in the curve is visible is a good indicator of the eps value to choose for the given k = minPts.
`plotCluster`	Boolean indicating if clusters identified with average linkage clustering should be visualized in PCs.

Performs clustering with dbscan (Density-based spatial clustering of applications with noise) for all SNPs in a window after filtering (i.e. initialize_SNP_time_series) and visualizes found clusters based on principal components. NOTE: The clusters identified here are not nessesarily identical with the clusters identified with average linkage clustering. Therefore if haplotype reconstruction was done using average linkage clustering the clusters shown here can be different from clusters of identified haplotype blocks. This method is rather indendet for inspection and getting an idea if haplotype reconstruction should rather be run using dbscan instead of average linkage clustering.

This package used the dbscan implementation of the package dbscan (Michael Hahsler) originally described by Ester et al. (1996).

Susanne U. Franssen

hbr ex_dat summary.hbr plot.hbr plot_cluster_trajectories plot_marker_trajectories map rev_map markers plot_hbr_freq inspect_window inspect_window_PCA inspect_window_avLink number_hbr