A practical framework and online tool for mutational signature analyses show intertissue variation and driver dependencies

Degasperi, Andrea; Amarante, Tauanne Dias; Czarnecki, Jan; Shooter, Scott; Zou, Xueqing; Glodzik, Dominik; Morganella, Sandro; Nanda, Arjun Scott; Badja, Cherif; Koh, Ching Chiek; Momen, Sophie; Georgakopoulos-Soares, Ilias; Dias, João M. L.; Young, Jamie; Memari, Yasin; Davies, Helen

doi:10.1038/s43018-020-0027-5

Cited by 181 publications

(224 citation statements)

References 32 publications

Supporting

Mentioning

216

Contrasting

Order By: Relevance

“…We note that CHORD performs similarly to HRDetect based on predictions on the BRCA-EU dataset (AUROC = 0.98 for both models) 9 . In addition, the predictions of CHORD and HRDetect on the PCAWG dataset 10 were concordant for the vast majority of samples (1506/1526; 99%) (Supplementary Fig. 10 ).…”

Section: Resultsmentioning

confidence: 75%

“…These features were used to develop a breast cancer-specific predictor of HRD known as HRDetect 9 . Application of this tool in primary tumors revealed that the prevalence of HRD extends beyond BRCA1/2 -deficient breast cancer tumors, and occurs at varying frequencies in different cancer types 10 . However, HRD rates in advanced metastatic cancer remain unclear, although these are the patients that are increasingly targeted with personalized treatments including PARP inhibitors for BRCA -deficiency 5 .…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Pan-cancer landscape of homologous recombination deficiency

et al. 2020

View full text Add to dashboard Cite

Homologous recombination deficiency (HRD) results in impaired double strand break repair and is a frequent driver of tumorigenesis. Here, we develop a genome-wide mutational scar-based pan-cancer Classifier of HOmologous Recombination Deficiency (CHORD) that can discriminate BRCA1- and BRCA2-subtypes. Analysis of a metastatic (n = 3,504) and primary (n = 1,854) pan-cancer cohort reveals that HRD is most frequent in ovarian and breast cancer, followed by pancreatic and prostate cancer. We identify biallelic inactivation of BRCA1, BRCA2, RAD51C or PALB2 as the most common genetic cause of HRD, with RAD51C and PALB2 inactivation resulting in BRCA2-type HRD. We find that while the specific genetic cause of HRD is cancer type specific, biallelic inactivation is predominantly associated with loss-of-heterozygosity (LOH), with increased contribution of deep deletions in prostate cancer. Our results demonstrate the value of pan-cancer genomics-based HRD testing and its potential diagnostic value for patient stratification towards treatment with e.g. poly ADP-ribose polymerase inhibitors (PARPi).

show abstract

Section: Resultsmentioning

confidence: 75%

Section: Introductionmentioning

confidence: 99%

Pan-cancer landscape of homologous recombination deficiency

et al. 2020

View full text Add to dashboard Cite

show abstract

“…It was suggested that SBS6, SBS14, SBS15, SBS20, SBS21, SBS26, and SBS44 were MMR-deficiency related 6 . In an independent analytical exercise, only two MMRassociated signatures were identified 24 , although variations of the signatures were seen in different tissue types 24 . An experimental process would help to obtain clarity in this regard [8][9][10][11] .…”

Section: Gene-specific Characteristics Of Mutational Signatures Of MMmentioning

confidence: 97%

“…Furthermore, gene-specific MMR signatures were seen in the International Cancer Genome Consortium (ICGC) cohort of >2,500 primary WGS cancers 24 . Indeed, biallelic MSH2/MSH6/MLH1 mutant tumors carried the same signature (RefSig MMR1) as ∆MSH2 /∆MSH6/∆MLH1 clones ( Figure 4B).…”

Section: Gene-specific Characteristics Of Mutational Signatures Of MMmentioning

confidence: 99%

Dissecting mutational mechanisms underpinning signatures caused by replication errors and endogenous DNA damage

Zou

Koh

Nanda

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Mutational signatures are imprints of pathophysiological processes arising through tumorigenesis. Here, we generate isogenic CRISPR-Cas9 knockouts (Δ) of 43 genes in human induced pluripotent stem cells, culture them in the absence of added DNA damage, and perform wholegenome sequencing of 173 daughter subclones. ΔOGG1, ΔUNG, ΔEXO1, ΔRNF168, ΔMLH1, ΔMSH2, ΔMSH6, ΔPMS1, and ΔPMS2 produce marked mutational signatures indicative of being critical mitigators of endogenous DNA changes. Detailed analyses reveal that 8-oxo-dG removal by different repair proteins is sequence-context-specific while uracil clearance is sequencecontext-independent. Signatures of mismatch repair (MMR) deficiency show components of C>A transversions due to oxidative damage, T>C and C>T transitions due to differential misincorporation by replicative polymerases, and T>A transversions for which we propose a ‘reverse template slippage’ model. ΔMLH1, ΔMSH6, and ΔMSH2 signatures are similar to each other but distinct from ΔPMS2. We validate these gene-specificities in cells from patients with Constitutive Mismatch Repair Deficiency Syndrome. Based on these experimental insights, we develop a classifier, MMRDetect, for improved clinical detection of MMR-deficient tumors.

show abstract

“…Based on the analysis of a large number of spectra, the "extraction and attribution" approach focuses on simultaneously finding de-novo signatures and exposures, usually implemented as non-negative matrix factorization algorithms. [4][5][6] The "attributiononly" approach, in contrast, assumes an existing compendium and aims to find the exposures that optimally reconstruct the observed spectra using standard constrained linear optimization methods. 7,8 De-novo signature discovery is a challenging task with several known issues, among them: defining the optimal number of signatures, discovering weaker signatures in a background of stronger ones, and resolving spill-over (bleeding) between signatures.…”

Section: Introductionmentioning

confidence: 99%

mutSigMapper: an R package to map spectra to mutational signatures based on shot-noise modeling

Candia

2020

Preprint

View full text Add to dashboard Cite

Summary: mutSigMapper aims to resolve a critical shortcoming of existing software for mutational signature analysis, namely that of finding parsimonious and biologically plausible exposures. By implementing a shot-noise-based model to generate spectral ensembles, this package addresses this gap and provides a quantitative, non-parametric assessment of statistical significance for the association between mutational signatures and observed spectra. Availability and implementation: The mutSigMapper R package is available under GPLv3 license at https://github.com/juliancandia/mutSigMapper. Its documentation provides additional details and demonstrates applications to biological datasets.

show abstract

A practical framework and online tool for mutational signature analyses show intertissue variation and driver dependencies

Cited by 181 publications

References 32 publications

Pan-cancer landscape of homologous recombination deficiency

Pan-cancer landscape of homologous recombination deficiency

Dissecting mutational mechanisms underpinning signatures caused by replication errors and endogenous DNA damage

mutSigMapper: an R package to map spectra to mutational signatures based on shot-noise modeling

Contact Info

Product

Resources

About