Skip to contents

Process read count matrix and segmentation

Source: R/cnaCalling.R
preProcSample2.Rd

Function adapted from facets::preProcSample() to process read count matrix and generates a segmentation tree. The modifications from the original function includes:

  • Incorporation of cluster and signal columns into the final result.

  • Change in the log ratio (cnlr) for allelic data, it is computed as the log ratio (cnlr) mean between the previous and next coverage positions.

Usage

preProcSample2(
  rcmat,
  het.thresh = 0.25,
  snp.nbhd = 250,
  cval = 25,
  gbuild = "hg38",
  hetscale = TRUE,
  ndepth = 5,
  ndepthmax = 1000
)

Source

This function is derived from the facets::preProcSample() function, with the modifications to fit its use in muscadet.

Seshan VE, Shen R (2021). facets: Cellular Fraction and Copy Numbers from Tumor Sequencing. R package version 0.6.2, https://github.com/mskcc/facets.

Arguments

rcmat

A data frame with 8 required columns: Chrom, Pos, NOR.DP, NOR.RD, TUM.DP, TUM.RD, cluster, and signal (data.frame).

het.thresh

VAF (Variant Allele Frequency) threshold to call variant positions heterozygous (numeric). Default: 0.25.

snp.nbhd

Window size for selecting loci to reduce serial correlation (numeric). Default: 250.

cval

Critical value for segmentation (numeric). Default: 25.

gbuild

Genome build used for alignment. One of "hg19", "hg38", or "mm10" (character). Default: "hg19".

hetscale

Logical value indicating whether log odds ratio (logOR) should be scaled to give more weight in the test statistics for segmentation and clustering (logical). Usually only 10% of snps are hets and hetscale gives the logOR contribution to T-square as 0.25/proportion of hets. Default: TRUE.

ndepth

Minimum depth in normal reference to keep (numeric). Default: 5.

ndepthmax

Maximum normal coverage threshold for filtering loci (numeric). Default: 1000.

Value

A list containing:

pmat

Read counts and other elements of all the loci.

gbuild

Genome build used for the analysis.

nX

Chromosome number for X (e.g., 23 for human, 20 for mouse).

clusters

Unique clusters from the processed data.

seg.tree

Segmentation tree for each chromosome.

jointseg

Segmented variant positions data.

hscl

Scaling factor for logOR data.

Details

The function processes variant positions data to generate a segmentation tree. It uses procSnps to compute initial values, adjusts the log ratio for allelic signals, and computes segmentation using segsnps.

SNPs (or variants) in a genome are not evenly spaced, and loci are sampled within the specified window (snp.nbhd) to reduce serial correlation.

References

facets-package package

Shen R, Seshan VE. FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 2016 Sep 19;44(16):e131. doi: 10.1093/nar/gkw520. PMID: 27270079; PMCID: PMC5027494.

See also

preProcSample, procSnps, segsnps

Examples

library("facets")

# Load example muscadet object
data(muscadet_obj)

counts <- muscadet_obj$cnacalling$combined.counts
counts <- counts[complete.cases(counts),]
counts_clus <- counts[which(counts$cluster == 1),]
result <- preProcSample2(counts_clus)