In sztup/scarHRD: Estimating Genomic scars signatures (LST, TAI,LOH)
Introduction
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.
Getting started
Minimum requirements
- Software: R
- Operating system: Linux, OS X, Windows
- R version: 3.4.0
Installation
scarHRD can be installed via devtools from github:
library(devtools)
install_github('sztup/scarHRD',build_vignettes = TRUE)
Citation
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.
Workflow overview
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 [@pmid25319062] http://www.cbs.dtu.dk/biotools/sequenza/ for copy number segmentation, Other tools (e.g. ASCAT [@pmid20837533]) may also be used in this step.
-
Determine the scar scores with scarHRD R package
Input file examples
The scarHRD input may be a detailed segmentation file from Sequenza:
example1<-system.file("extdata", "test1.small.seqz.gz", package = "scarHRD")
a<-read.table(example1, header=T)
head(a)
or a simplified file, including the total, and allele-specific copy-number:
example2<-system.file("extdata", "test1.small.seqz.gz", package = "scarHRD")
a<-read.table(example2, header=T)
head(a)
Usage example
scar_score("/examples/test1.small.seqz.gz",reference = "grch38", seqz=TRUE)
scar_score("/examples/test2.txt",reference = "grch38", seqz=FALSE)
If your file does not contain 'chr' in chromosome names, use chr.in.names=FALSE.
Optional parameters
reference -- the reference genome used, grch38 or grch37 or mouse
chr.in.names -- default: TRUE, set to FALSE if input file does not contain 'chr' in chromosome names
Genomic scar scores
Loss of Heterozygosity (HRD-LOH)
The HRD-LOH score was described based on investigation in SNP-array-based copy number profiles of ovarian cancer [@pmid22933060]. 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.
## Large Scale Transitions (LST)
The presence of Large Scale Transitions in connection with homologous recombination deficiency was first in studied in basal-like breast cancer [@pmid23047548]. Based on SNP-array derived copy number profiles ... the number of telomeric allelic imbalances
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...
## Number of Telomeric Allelic Imbalances
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 [@pmid22576213].
The number of telomeric allelic imbalances is the number AIs that extend to the telomeric end of a chromosome..
# Running on mouse genomes
`scarHRD()` can be run on mouse genome using `referrence="mouse"`. However it is problematic to evaluate the HRD-score in mouse. The mouse chromosome are telocentric, meaning that the centrome is next to the telomeric region. This makes it hard to determine and evaluate Telomeric Allelic Imbalance score. Moreover the mouse cancer models are usually isogenic, which yield low number of heterogeneous positions, which are essential for determining HRD-LOH scores. So calculating genomic scar scores (telomeric AI, HRD-LOH, LST) is possible for mice, but it has a very limited utility.
# References
sztup/scarHRD documentation built on Aug. 26, 2020, 4:19 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.
Introduction
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.
Getting started
Minimum requirements
- Software: R
- Operating system: Linux, OS X, Windows
- R version: 3.4.0
Installation
scarHRD can be installed via devtools from github:
library(devtools) install_github('sztup/scarHRD',build_vignettes = TRUE)
Citation
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.
Workflow overview
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 [@pmid25319062] http://www.cbs.dtu.dk/biotools/sequenza/ for copy number segmentation, Other tools (e.g. ASCAT [@pmid20837533]) may also be used in this step.
-
Determine the scar scores with scarHRD R package
Input file examples
The scarHRD input may be a detailed segmentation file from Sequenza:
example1<-system.file("extdata", "test1.small.seqz.gz", package = "scarHRD") a<-read.table(example1, header=T) head(a)
or a simplified file, including the total, and allele-specific copy-number:
example2<-system.file("extdata", "test1.small.seqz.gz", package = "scarHRD") a<-read.table(example2, header=T) head(a)
Usage example
scar_score("/examples/test1.small.seqz.gz",reference = "grch38", seqz=TRUE) scar_score("/examples/test2.txt",reference = "grch38", seqz=FALSE)
If your file does not contain 'chr' in chromosome names, use chr.in.names=FALSE.
Optional parameters
reference -- the reference genome used, grch38 or grch37 or mouse
chr.in.names -- default: TRUE, set to FALSE if input file does not contain 'chr' in chromosome names
Genomic scar scores
Loss of Heterozygosity (HRD-LOH)
The HRD-LOH score was described based on investigation in SNP-array-based copy number profiles of ovarian cancer [@pmid22933060]. 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.
## Large Scale Transitions (LST) The presence of Large Scale Transitions in connection with homologous recombination deficiency was first in studied in basal-like breast cancer [@pmid23047548]. Based on SNP-array derived copy number profiles ... the number of telomeric allelic imbalances
## Number of Telomeric Allelic Imbalances 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 [@pmid22576213].
# Running on mouse genomes `scarHRD()` can be run on mouse genome using `referrence="mouse"`. However it is problematic to evaluate the HRD-score in mouse. The mouse chromosome are telocentric, meaning that the centrome is next to the telomeric region. This makes it hard to determine and evaluate Telomeric Allelic Imbalance score. Moreover the mouse cancer models are usually isogenic, which yield low number of heterogeneous positions, which are essential for determining HRD-LOH scores. So calculating genomic scar scores (telomeric AI, HRD-LOH, LST) is possible for mice, but it has a very limited utility. # References
sztup/scarHRD documentation built on Aug. 26, 2020, 4:19 a.m.
## Large Scale Transitions (LST) The presence of Large Scale Transitions in connection with homologous recombination deficiency was first in studied in basal-like breast cancer [@pmid23047548]. Based on SNP-array derived copy number profiles ... the number of telomeric allelic imbalances
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...
## Number of Telomeric Allelic Imbalances 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 [@pmid22576213].
The number of telomeric allelic imbalances is the number AIs that extend to the telomeric end of a chromosome..
# Running on mouse genomes `scarHRD()` can be run on mouse genome using `referrence="mouse"`. However it is problematic to evaluate the HRD-score in mouse. The mouse chromosome are telocentric, meaning that the centrome is next to the telomeric region. This makes it hard to determine and evaluate Telomeric Allelic Imbalance score. Moreover the mouse cancer models are usually isogenic, which yield low number of heterogeneous positions, which are essential for determining HRD-LOH scores. So calculating genomic scar scores (telomeric AI, HRD-LOH, LST) is possible for mice, but it has a very limited utility. # References
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.