Contents

1 Introduction

To evaluate the aneuploidy and prevalence of clonal or quasiclonal tumors, we developed a novel tool to characterize the mosaic tumor genome on the basis of one major assumption: the genome of a heterogeneous multi-cell tumor biopsy can be sliced into a chain of segments that are characterized by homogeneous somatic copy number alternations (SCNAs) and B allele frequencies (BAFs). The model, termed BubbleTree, utilizes both SCNAs and the interconnected BAFs as markers of tumor clones to extract tumor clonality estimates. BubbleTree is an intuitive and powerful approach to jointly identify ASCN, tumor purity and (sub)clonality, which aims to improve upon current methods to characterize the tumor karyotypes and ultimately better inform cancer diagnosis, prognosis and treatment decisions.

2 Quickstart to Using BubbleTree

To perform a BubbleTree analysis, data pertaining to the position and B allele frequency of heterozygous snps in the tumor sample and segmented copy number information including the position, number of markers/segment and log2 copy number ratio between tumor and normal samples must first be obtained.

2.1 Preparing Data for BubbleTree

BubbleTree was developed using both whole exome sequencing (WES) and whole genome sequening (WGS) NGS data from paired tumor/normal biopsies, but this model should also be applicable to array comparative genomic hybridization (aCGH) and single nucleotide polymorphism (SNP) array data.

There are many methods for generating and processing sequencing data in preparation for BubbleTree analysis. In the following section we provide example workflows starting from WES NGS which can be adapted as needed to alternate inputs.

2.2 Preparing Sequence Variation Data

The primary BubbleTree requirement for sequence variant information is a GRanges object containing the B alelle frequencies and genomic positions of variants known to be heterozygous in the paired normal sample.

Mapped reads from tumor and normal tissue can be processed with mutation caller software such as VarScan or MUTECT. In this example, we use a hypothetical vcf file from VarScan output which contains mutation calls from both normal and tumor samples.

2.2.1 Preparing BAF Data From VarScan

Assume that you have loaded the data snp.dat like this:

head(snp.dat)

  CHROM    POS  ID REF ALT QUAL FILTER LT.rna.dp LN.rna.dp ON.rna.dp OT.rna.dp BT.wes.dp LT.wes.dp LN.wes.dp ON.wes.dp
1  chr1  54757 rs202000650   T   G    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
2  chr1 564636           .   C   T    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
3  chr1 564862   rs1988726   T   C    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
4  chr1 564868   rs1832728   T   C    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
5  chr1 565454   rs7349151   T   C    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
6  chr1 565464   rs6594030   T   C    .   PASS        NA        NA        NA        NA        NA        NA        NA        NA
  OT.wes.dp LT.wgs.dp LN.wgs.dp ON.wgs.dp OT.wgs.dp LT.rna.freq LN.rna.freq ON.rna.freq OT.rna.freq BT.wes.freq LT.wes.freq
1        NA        25        24        27        19          NA          NA          NA          NA          NA          NA
2        NA        21        NA        NA        14          NA          NA          NA          NA          NA          NA
3        NA        10        15        55        13          NA          NA          NA          NA          NA          NA
4        NA        10        12        60        14          NA          NA          NA          NA          NA          NA
5        NA        21        14        26        24          NA          NA          NA          NA          NA          NA
6        NA        25        16        33        29          NA          NA          NA          NA          NA          NA
  LN.wes.freq ON.wes.freq OT.wes.freq LT.wgs.freq LN.wgs.freq ON.wgs.freq OT.wgs.freq
1          NA          NA          NA      0.2400      0.1667      0.2222      0.3684
2          NA          NA          NA      0.0000          NA          NA      0.1429
3          NA          NA          NA      0.4000      0.5333      0.9091      0.7692
4          NA          NA          NA      0.5000      0.6667      0.9333      0.7857
5          NA          NA          NA      0.1429      0.3571      0.6538      0.6250
6          NA          NA          NA      0.2000      0.3750      0.7273      0.5862

Identify the germline heterozygous loci:

is.hetero <- function(x, a=0.3, b=0.7) {
 (x - a)  *  (b - x) >= 0
}

wgs.snp.ss <- subset(snp.dat, ! CHROM %in% c("chrX", "chrY") & 
                         LN.wgs.dp >= 15 & 
                         ON.wgs.dp >=15 & 
                         is.hetero(LN.wgs.freq, 0.4, 0.6) & 
                         is.hetero(ON.wgs.freq, 0.4, 0.6))

Then convert to the GRanges object:

library(GenomicRanges)
wgs.hetero.grs <- list()
wgs.hetero.grs$lung <- with(wgs.snp.ss, GRanges(CHROM, IRanges(POS, POS), mcols=cbind(LT.wgs.dp, LT.wgs.freq)))
wgs.hetero.grs$ovary <- with(wgs.snp.ss, GRanges(CHROM, IRanges(POS, POS), mcols=cbind(OT.wgs.dp, OT.wgs.freq)))
names(elementMetadata(wgs.hetero.grs$lung)) <- names(elementMetadata(wgs.hetero.grs$ovary))  <- c("dp", "freq")

The B-allele frequency data is extracted using the Bioconductor package VariantAnnotation and converted from string to numeric format.

2.2.2 Preparing CNV Data from DNAcopy

The object seg is the segment call output from DNAcopy and min.num here specifies the minimum segment size to keep

library(GenomicRanges)
min.num <- 10
cnv.gr <- with(subset(seg$output, num.mark >= min.num & ! chrom %in% c("chrX", "chrY")) , GRanges(chrom, IRanges(loc.start, loc.end), mcols=cbind(num.mark, seg.mean)))

Then merge the SNP and CNV GRanges objects.

Example data in the desired format is provided as part of this package as GRanges objects and can be loaded as shown below. To utilize this vignette, you must first load BubbleTree below. You don’t need to use “suppressMessages”.

suppressMessages(
    library(BubbleTree)
)

allCall.lst is pre-calculated CNV data. allRBD.lst is simply the RBD data from below.

load(system.file("data", "allCall.lst.RData", package="BubbleTree"))
head(allCall.lst[[1]]@rbd)
##   seqnames    start       end     width strand seg.id num.mark    lrr
## 1    chr10    93890  38769716  38675827      *    806    31699 0.1413
## 2    chr10 38877329 135523936  96646608      *    808    74425 0.1415
## 3    chr11   133952 134946370 134812419      *    812   102934 0.1412
## 4    chr12    60000 133841793 133781794      *    813   103392 0.1413
## 5    chr13 19020000 115109861  96089862      *    814    76080 0.1419
## 6    chr14 20191636 107288640  87097005      *    823    68709 0.1425
##       kurtosis        hds     hds.sd het.cnt seg.size
## 1 -0.093044958 0.01851852 0.06051370   10400 1.487216
## 2 -0.048701274 0.01851852 0.06046402   25414 3.491784
## 3 -0.018840021 0.01851852 0.06052843   36183 4.829335
## 4 -0.007149860 0.01851852 0.06096056   36798 4.850823
## 5 -0.022536940 0.01851852 0.06057517   27278 3.569431
## 6 -0.004935614 0.01851852 0.06166893   25001 3.223607

However, if you wish to create your own RBD object from your input, you would use the code below. There is test data available in this package that is used for demonstration purposes.

# load sample files
load(system.file("data", "cnv.gr.rda", package="BubbleTree"))
load(system.file("data", "snp.gr.rda", package="BubbleTree"))

# load annotations
load(system.file("data", "centromere.dat.rda", package="BubbleTree"))
load(system.file("data", "cyto.gr.rda", package="BubbleTree"))
load(system.file("data", "cancer.genes.minus2.rda", package="BubbleTree"))
load(system.file("data", "vol.genes.rda", package="BubbleTree"))
load(system.file("data", "gene.uni.clean.gr.rda", package="BubbleTree"))


# initialize RBD object
r <- new("RBD", unimodal.kurtosis=-0.1)

# create new RBD object with GenomicRanges objects for SNPs and CNVs
rbd <- makeRBD(r, snp.gr, cnv.gr)
head(rbd)
##   seqnames    start      end    width strand seg.id num.mark     lrr
## 1     chr1    65625  2066855  2001231      *      1      548  0.1997
## 2     chr1  2075796 38489397 36413602      *      2     5284 -0.4146
## 3     chr1 38511244 39761601  1250358      *      3       72 -0.0511
## 4     chr1 39763396 39982177   218782      *      4      112  0.0401
## 5     chr1 39988109 43905367  3917259      *      5      601  0.0372
## 6     chr1 43905709 44128685   222977      *      6       53  0.2822
##     kurtosis     hds     hds.sd het.cnt   seg.size
## 1  0.1830119 0.01220 0.05007095      71 0.24809178
## 2 -1.9020390 0.35600 0.06296830     501 2.39218420
## 3 -2.0000000 0.15555 0.03471894       2 0.03259600
## 4 -0.9912620 0.12060 0.06761821       4 0.05070489
## 5 -1.4809979 0.14560 0.06186714      47 0.27208605
## 6 -2.0000000 0.07280 0.05176022       2 0.02399428
# create a new prediction
btreepredictor <- new("BTreePredictor", rbd=rbd, max.ploidy=6, prev.grid=seq(0.2,1, by=0.01))
pred <- btpredict(btreepredictor)

# create rbd plot
btreeplotter <- new("BTreePlotter", max.ploidy=5, max.size=10)
btree <- drawBTree(btreeplotter, pred@rbd)
## Warning in drawBTree(btreeplotter, pred@rbd): More ploidy might be suggested: 1.6, 1.6, 2.2, 1.7, 1.9, 1.5, 1.7, 2.2, 1.9, 2.1, 1.7
print(btree)