07-generateMixtures: Generating sets of artificially mixed and altered...

In CloneData: Data to Support CloneSeeker Algorithm

Generating sets of artificially mixed and altered heterogeneous data

Description

Generating and saving a 'simulated' tumor data set by artificially mixing and altering real SNP array data that can be used in clonal heterogeneity analysis to assess accuracy of algorithms.

Usage

generateMixtures(dataPath, mixPath, nPerK, segmentedData, ID_pool, pos)

Arguments

dataPath	path to which simulated tumors will be saved.
mixPath	path to which artificially mixed and altered SNP array data will be saved.
nPerK	a vector of integers denoting the number of tumors to generate for each possible number of clones, where the nth entry dictates how many n-clone tumors will be generated.
segmentedData	segmented SNP array data from which mixtures will be generated; must contain following columns: 'loc.start' (segment start locus), 'loc.end' (segment end locus), 'seg.median' (median Log R ratio), 'SamID' (sample ID), 'chrom' (chromosome number), 'AvgBAF' (average B allele fraction for segment), 'num.mark' (number of markers per segment).
ID_pool	a list of sample IDs from segmentedData from which samples will be drawn to generate artificial mixtures.
pos	a data frame with two columns, Chr and Position, defining the chromsoomal locations of the simulated SNPs.

Details

A set of artificial mixtures (with CNVs artificially added) can be generated from real SNP array data. The number of artificial mixtures to generate - and how many mixtures for each possible number of clones to generate - can be set with the input parameters.

Value

The generateMixtures function generates and saves two lists for each mixture: a 'tumor' (consisingt of artificially altered real data making up the 'clones' of the mixture, saved in the path 'simpath'), with objects: psi, a vector of clonal fractions, clones, which is a list of tumor clones, each of which in turn consists of a data frame cn and a data frame seq, a list altered (a list of segments artificially altered), and a list change (the copy number change introduced to the altered segments); and a simulated data object (saved in the path 'datapath'), with objects: cn.data and se .data. Each component is itself a data frame. Note that in some cases, one of these data frames may have zero rows or may be returned as an NA.

Each list in the cn component contains seven columns:

chr

the chromosome number;

start

the starting locus of each genomic segment;

end

the ending locus of each genomic segment;

A

the first allelic copy number at each genomic segment;

B

the second allelic copy number at each genomic segment;

seg

the segment number; and

parent.index

the index of the clone from which this clone is descended (equals 0 if the clone is an original tumor clone).

Each list in the seq component contains seven columns:

chr

the chromosome number;

start

the locus of the simulated SNVs;

seg

the segment on which each SNV occurs;

mut.id

the id unique id number for each simulated SNV;

mutated.copies

the number of copies of the mutated allele at each SNV;

alllele

which allele (A or B) is mutated at each SNV; and

normal.copies

the number of copies of the unmutated allele at each SNV.

The cn.data component contains seven columns:

chr

the chromosome number;

seq

a unique segment identifier;

LRR

simulated segment-wise log ratios;

BAF

simulated segment-wise B allele frequencies;

X and Y

simulated intensities for two separate alleles/haplotypes per segment; and

markers

the simulated number of SNPS per segment.

The seq.data component contains eight columns:

chr

the chromosome number;

seq

a unique "segment" identifier;

mut.id

a unique mutation identifier;

refCounts and varCounts

the simulated numbers of reference and variant counts per mutation;

VAF

the simulated variant allele frequency;

totalCounts

the simulated total number of read counts; and

status

a character (that should probably be a factor) indicating whether a variant should be viewed as somatic or germline.

Author(s)

Kevin R. Coombes [e-mail], Mark Zucker [e-mail]

References

Zucker MR, Abruzzo LV, Herling CD, Barron LL, Keating MJ, Abrams ZB, Heerema N, Coombes KR. Inferring Clonal Heterogeneity in Cancer using SNP Arrays and Whole Genome Sequencing. Bioinformatics. To appear. doi: 10.1093/bioinformatics/btz057.

Examples

# Set of 300 simulated 'tumors' generated by artificially mixing and
# altering real data; 60 samples with one #clone, 60 with 2 clones,
# ..., 60 with 5 clones.
data("hapmapSegments", package = "CloneData")
data("snpPositions", package = "CloneData")
IDset <- c('NA07019', 'NA12234', 'NA12249', 'NA12753', 'NA12761',
           'NA18545', 'NA18975', 'NA18999', 'NA18517')
# Generating the data set:
## Not run: 
generateMixtures(dataPath = 'mixdat', mixPath = 'mixsim',
                 nPerK = rep(60,5),  segmentedData = hapmapSegments,
                 ID_pool = IDset, pos = snpPositions)

## End(Not run)

CloneData documentation built on July 1, 2022, 3 a.m.