call_events: Call CNV events from a set of read counts
In malemay/delgbs: Detection of deletions and duplications from GBS data

call_events

R Documentation

Call CNV events from a set of read counts

Description

This function performs all the steps needed to go from a data.frame read counts per bin and per individual to a set of CNVs. The steps involved are the normalization of the read counts per sample, the removal of bins deemed uninformative for the purposes of CNV calling, the computation of log2 ratios and segmentation based on these ratios, and finally CNV calling based on user-defined log2 ratio thresholds.

Usage

call_events(read_counts, min_reads, max_reads, seg_threshold, homdel = -2.5,
  hetdel = -0.5, dup = 0.2)

Arguments

`read_counts`	a `data.frame` of read counts as generated from function `link{read_counts}`.
`min_reads`	a single numeric value. The minimum mean number of reads per individual that a bin must have to be kept for further processing.
`max_reads`	a single numeric value. The maximum mean number of reads per individual that a bin can have to be kept for further processing.
`seg_threshold`	a single numeric value. The threshold parameter used for segmentation. See function `segment` for more details.
`homdel`	a single numeric value. The threshold to be used for calling homozygous deletions; all segments with a mean log2 ratio less than this values is considered called as a homozygous deletion.
`hetdel`	a single numeric value. The threshold to be used for calling hemizygous deletions; all segments with a mean log2 ratio less than this value and great than or equal to `homdel` is called as a hemizygous deletion.
`dup`	a single numeric value. The threshold to be used for calling duplications; all segments with a mean log2 ratio larger than this value is called as a duplication.

Value

a list of two elements : - segdels: a data.frame containing the CNV events called and some metadata about them. - log2_ratios: a data.frame of log2 ratios per bin and per sample. The first three columns indicate the genomic position of the bin, whereas each remaining column represents the data for a single sample (column names correspond to sample names).