- Introduction
- Getting started
- Citation
- Workflow overview
- Genomic scar scores
- Running on mouse genomes
- References
scarHRD is an R package which determines the levels of homologous recombination deficiency (telomeric allelic imbalance, loss off heterozygosity, number of large-scale transitions) based on NGS (WES, WGS) data.
The first genomic scar based homologous recombination deficiency measures were produced using SNP arrays. Since this technology has been largely replaced by next generation sequencing it has become important to develop algorithms that derive the same type of genomic scar-scores from next generation sequencing (WXS, WGS) data. In order to perform this analysis, here we introduce the scarHRD R package and show that using this method the SNP-array based and next generation sequencing based derivation of HRD scores show good correlation.
Zsofia Sztupinszki
Boston Children's Hospital
contact: [email protected]
- Software: R
- Operating system: Linux, OS X, Windows
- R version: 3.5.0
- latest version of Sequenza R package must be installed from Bitbucket https://bitbucket.org/sequenza_tools/sequenza
library(devtools)
install_bitbucket('sequenza_tools/sequenza')scarHRD can be installed via devtools from github:
library(devtools)
install_github('sztup/scarHRD',build_vignettes = TRUE)A modification of the copynumber R package needs to be used which can be installed via devtools from github:
library(devtools)
install_github('aroneklund/copynumber')Please cite the following paper: Sztupinszki et al, Migrating the SNP array-based homologous recombination deficiency measures to next generation sequencing data of breast cancer, npj Breast Cancer, https://www.nature.com/articles/s41523-018-0066-6.
A typical workflow of determining the genomic scar scores for a tumor sample has the following steps:
- Call allele specific copy number profile on paired normal-tumor BAM files. This step has to be executed before running scarHRD. We recommend using Sequenza (Favero et al. 2015) http://www.cbs.dtu.dk/biotools/sequenza/ for copy number segmentation, Other tools (e.g. ASCAT (Van Loo et al. 2010)) may also be used in this step.
This step is time-consuming and compute-intensive.
Example for using Sequenza:
sequenza-utils bam2seqz -gc /reference/GRCh38.gc50Base.txt.gz --fasta /reference/GRCh38.d1.vd1.fa -n /data/normal.bam --tumor /data/tumor.bam -C chr1 chr2 chr3 chr4 chr5 chr6 chr7 chr8 chr9 chr10 chr11 chr12 chr13 chr14 chr15 chr16 chr17 chr18 chr19 chr20 chr21 chr22 chr23 chr24 chrX | sequenza-utils seqz_binning -w 50 -s - | gzip > /results/tumor_small.seqz.gz Further details can be found in the Vignette of Sequenza: https://cran.r-project.org/web/packages/sequenza/vignettes/sequenza.pdf
- Determine the scar scores with scarHRD R package.
This step only takes a few minutes.
The scarHRD input may be a detailed segmentation file from Sequenza, in case there is a reliable estimation of ploidy of the tumor sample is known, it should be sumbitted in the ploidy argument of the scarHRD function, otherwise ploidy between 1 and 5.5 will be tested:
a<-read.table("/examples/test1.small.seqz.gz", header=T)
head(a)## chromosome position base.ref depth.normal depth.tumor depth.ratio Af
## 1 chr1 12975 N 7 20 2.841 1.000
## 2 chr1 13020 A 8 28 3.500 0.964
## 3 chr1 13026 N 15 43 2.964 1.000
## 4 chr1 13038 T 11 35 3.182 0.971
## 5 chr1 13041 A 11 37 3.364 0.946
## 6 chr1 13077 N 26 65 2.465 1.000
## Bf zygosity.normal GC.percent good.reads AB.normal AB.tumor tumor.strand
## 1 0 hom 60 51 N . 0
## 2 0 hom 60 28 A G0.036 G1.0
## 3 0 hom 59 51 N . 0
## 4 0 hom 59 35 T C0.029 C1.0
## 5 0 hom 59 37 A G0.054 G0.5
## 6 0 hom 62 51 N . 0
or a simplified file, including the total, and allele-specific copy-number:
a<-read.table("/examples/test2.txt", header=T)
head(a)## SampleID Chromosome Start_position End_position total_cn A_cn B_cn
## 1 SamplePatient1 chr1 14574 952448 5 0 5
## 2 SamplePatient1 chr1 953394 1259701 3 0 3
## 3 SamplePatient1 chr1 1278085 4551743 2 0 2
## 4 SamplePatient1 chr1 4551885 14124232 2 0 2
## 5 SamplePatient1 chr1 14161231 31062374 3 1 2
## 6 SamplePatient1 chr1 31074785 47428120 4 2 2
## ploidy
## 1 3.7
## 2 3.7
## 3 3.7
## 4 3.7
## 5 3.7
## 6 3.7
library("scarHRD")
scar_score("F:/Documents/scarHRD/examples/test1.small.seqz.gz",reference = "grch38", seqz=TRUE)## Preprocessing started...
## Processing chr1: 18 variant calls; 6290 heterozygous positions; 549112 homozygous positions.
## Processing chr2: 22 variant calls; 4934 heterozygous positions; 394216 homozygous positions.
##
|=================================================================| 100%
## Preprocessing finished
## Determining HRD-LOH, LST, TAI
## HRD Telomeric AI LST HRD-sum
## [1,] 1 2 0 3
scar_score("F:/Documents/scarHRD/examples/test2.txt",reference = "grch38", seqz=FALSE)## Determining HRD-LOH, LST, TAI
## HRD Telomeric AI LST HRD-sum
## [1,] 25 35 33 93
seg -- input file name
reference -- the reference genome used, grch38 or grch37 or mouse (default: grch38)
seqz -- TRUE if the input file is a smallo.seqz.gz file, otherwise FALSE (default: TRUE)
ploidy -- optional, previously estimated ploidy of the sample outputdir -- optional, the path to the output directory
chr.in.names -- optional, default: TRUE, set to FALSE if input file does not contain 'chr' in chromosome names.
The HRD-LOH score was described based on investigation in SNP-array-based copy number profiles of ovarian cancer (Abkevich et al. 2012). In this paper the authors showed that the samples with deficient BRCA1, BRCA2 have higher HRD-LOH scores compared to BRCA-intact samples, thus this measurement may be a reliable tool to estimate the sample's homologous recombination capacity.
The definition of a sample's HRD-LOH score is the number of 15 Mb exceeding LOH regions which do not cover the whole chromosome.
In the first paper publishing HRD-LOH-score (Abkevich et al., 2012) the authors examine the correlation between HRD-LOH-score and HR deficiency calculated for different LOH region length cut-offs. In that paper the cut-off of 15 Mb approximately in the middle of the interval was arbitrarily selected for further analysis. The authors argue that the rational for this selection rather than selecting the cut-off with the lowest p-value is that the latter cut-off is more sensitive to statistical noise present in the data.
In our manuscript we also investigated if this 15 Mb cutoff is appropriate for WXS-based HRD-LOH score.We followed the same principles as Abkievits et al, thus while there was small difference between the p-values for the different minimum length cutoff values, we chose to use the same, 15 Mb limit as Abkevich et al. We also performed Spearman rank correlation between the SNP-array-based and WXS-based HRD-LOH scores for the different cutoff minimum LOH length cutoff (manuscript, Supplementary Figure S3C). Here the 14 Mb and 15 Mb cutoff-based WXS-HRD-LOH score had the highest correlation with the SNP-based HRD score. (0.700 and 0.695 respectively). This result reassured our choice of using the 15 Mb cutoff like in the SNP-array-based HRD-LOH score.
The presence of Large Scale Transitions in connection with homologous recombination deficiency was first studied in basal-like breast cancer (Popova et al. 2012). Based on SNP-array derived copy number profiles BRCA1-inactivated cases had showed higher number of large scale transitions.
A large scale transition is defined as a chromosomal break between adjacent regions of at least 10 Mb, with a distance between them not larger than 3Mb.
Allelic imbalance (AI) is the unequal contribution of parental allele sequences with or without changes in the overall copy number of the region. Our group have previously found, that the number telomeric AIs is indicative of defective DNA repair in ovarian cancer and triple-negative breast cancer, and that higher number of telomeric AI is associated with better response to cisplatin treatment (Birkbak et al. 2012).
The number of telomeric allelic imbalances is the number AIs that extend to the telomeric end of a chromosome.
Abkevich, V., K. M. Timms, B. T. Hennessy, J. Potter, M. S. Carey, L. A. Meyer, K. Smith-McCune, et al. 2012. “Patterns of genomic loss of heterozygosity predict homologous recombination repair defects in epithelial ovarian cancer.” Br. J. Cancer 107 (10): 1776–82.
Birkbak, N. J., Z. C. Wang, J. Y. Kim, A. C. Eklund, Q. Li, R. Tian, C. Bowman-Colin, et al. 2012. “Telomeric allelic imbalance indicates defective DNA repair and sensitivity to DNA-damaging agents.” Cancer Discov 2 (4): 366–75.
Favero, F., T. Joshi, A. M. Marquard, N. J. Birkbak, M. Krzystanek, Q. Li, Z. Szallasi, and A. C. Eklund. 2015. “Sequenza: allele-specific copy number and mutation profiles from tumor sequencing data.” Ann. Oncol. 26 (1): 64–70.
Popova, T., E. Manie, G. Rieunier, V. Caux-Moncoutier, C. Tirapo, T. Dubois, O. Delattre, et al. 2012. “Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1/2 inactivation.” Cancer Res. 72 (21): 5454–62.
Van Loo, P., S. H. Nordgard, O. C. Lingj?rde, H. G. Russnes, I. H. Rye, W. Sun, V. J. Weigman, et al. 2010. “Allele-specific copy number analysis of tumors.” Proc. Natl. Acad. Sci. U.S.A. 107 (39): 16910–5.