reconstruct_hb-SNP_time_series-method: Reconstruction of haplotype-blocks

In haploReconstruct: Reconstruction of Haplotype-Blocks from Time Series Data

Description

This function reconstructs haplotype-blocks from a SNP_time_series object based on correlated allele frequency changes aacross time points and replicates

Usage

## S4 method for signature 'SNP_time_series'
reconstruct_hb(object, chrom, min.cl.size = 4,
  min.cl.cor = 0.8, min.inter = 2, single.win = F, transf = T,
  scaleSNP = T, pos.cor = F, clusterM = "avLink", eps = 0.3)

Arguments

object	object of the class SNP_time_series data
chrom	a string specifying the name of the chromosome present in the object for which haplotype-block should be reconstructed
min.cl.size	numeric specifying the minimum number of correlated markers in a window for haplotype-block reconstruction
min.cl.cor	numeric specifying the correlation between marker SNPs required using average linkage clustering for markers to be assembled in one cluster
min.inter	numeric specifying the minimum number of markers in two clusters of overlapping windows required to be identical for cluster elongation across windows to build haplotype-blocks
single.win	boolean specifying that are supported by only a cluster in one window are also included in the markers of reconstructed blocks. If FALSE only hbrs are returned that span at least two windows and only markers being present in the intersection between overlapping windows are included.
transf	boolean indicating if frequency data should be square root transformed prior to calaculation pariwise correlations with Pearson's correlation coefficient.
scaleSNP	boolean indicating whether time series allele frequency data for each SNP should be scaled (scaling to a mean of zero and standard deviation of 1.
pos.cor	boolean indicating if negative correlations should be set to zero prior to clustering.
clusterM	indicating the clustering method, choose from: "avLink" (average linkage clustering) and "dbscan".
eps	size of the epsilon neighborhood when clustering with dbscan. For details please refer to the dbscan package.

Details

This function reconstructs haplotype-blocks via markers that show correlated frequency changes across multiple time-points and replicates. In a sliding window based approach SNPs are clustered by average linkage clustering based on correlations of their frequencies across time points and replicates. Clusters are build for each window using the parameters min.cl.size and min.cl.cor. Clustered in overlaping windows (window size/2) are elongated to haplotype-blocks if two adjacent clusters have min.inter identical markers./n Reconstructed haplotype blocks were constructed from at least two overlapping windows and consist of markers that are present in two overlaping window clusters.

Value

an object of the class hbr

Author(s)

Susanne U. Franssen

References

Franssen et al. 2016, Reconstruction of haplotype-blocks selected during experimental evolution, MBE

See Also

ex_dat initialize_SNP_time_series SNP_time_series

haploReconstruct documentation built on May 2, 2019, 1:46 p.m.