Dataset Information

Hidden biases in germline structural variant detection.

ABSTRACT:

Background

Genomic structural variations (SV) are important determinants of genotypic and phenotypic changes in many organisms. However, the detection of SV from next-generation sequencing data remains challenging.

Results

In this study, DNA from a Chinese family quartet is sequenced at three different sequencing centers in triplicate. A total of 288 derivative data sets are generated utilizing different analysis pipelines and compared to identify sources of analytical variability. Mapping methods provide the major contribution to variability, followed by sequencing centers and replicates. Interestingly, SV supported by only one center or replicate often represent true positives with 47.02% and 45.44% overlapping the long-read SV call set, respectively. This is consistent with an overall higher false negative rate for SV calling in centers and replicates compared to mappers (15.72%). Finally, we observe that the SV calling variability also persists in a genotyping approach, indicating the impact of the underlying sequencing and preparation approaches.

Conclusions

This study provides the first detailed insights into the sources of variability in SV identification from next-generation sequencing and highlights remaining challenges in SV calling for large cohorts. We further give recommendations on how to reduce SV calling variability and the choice of alignment methodology.

SUBMITTER: Khayat MM

PROVIDER: S-EPMC8686633 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundLess than two percent of the human genome is protein coding, yet that small fraction harbours the majority of known disease causing mutations. Despite rapidly falling whole genome sequencing (WGS) costs, much research and increasingly the clinical use of sequence data is likely to remain focused on the protein coding exome. We set out to quantify and understand how WGS compares with the targeted capture and sequencing of the exome (exome-seq), for the specific purpose of identifying single nucleotide polymorphisms (SNPs) in exome targeted regions.ResultsWe have compared polymorphism detection sensitivity and systematic biases using a set of tissue samples that have been subject to both deep exome and whole genome sequencing. The scoring of detection sensitivity was based on sequence down sampling and reference to a set of gold-standard SNP calls for each sample. Despite evidence of incremental improvements in exome capture technology over time, whole genome sequencing has greater uniformity of sequence read coverage and reduced biases in the detection of non-reference alleles than exome-seq. Exome-seq achieves 95% SNP detection sensitivity at a mean on-target depth of 40 reads, whereas WGS only requires a mean of 14 reads. Known disease causing mutations are not biased towards easy or hard to sequence areas of the genome for either exome-seq or WGS.ConclusionsFrom an economic perspective, WGS is at parity with exome-seq for variant detection in the targeted coding regions. WGS offers benefits in uniformity of read coverage and more balanced allele ratio calls, both of which can in most cases be offset by deeper exome-seq, with the caveat that some exome-seq targets will never achieve sufficient mapped read depth for variant detection due to technical difficulties or probe failures. As WGS is intrinsically richer data that can provide insight into polymorphisms outside coding regions and reveal genomic rearrangements, it is likely to progressively replace exome-seq for many applications.

Project description:BackgroundTumor-only sequencing, implemented for the identification of somatic variants, is oftentimes used for the detection of actionable germline variants. We sought to determine whether tumor-only sequencing assays are suitable for detection of actionable germline variants, given their importance for the delivery of targeted therapies and risk-reducing measures.Patients and methodsThe detection of germline variants affecting moderate- and high-penetrance cancer susceptibility genes (CSGs) by tumor-only sequencing was compared to clinical germline testing in 21 333 cancer patients who underwent tumor and germline testing using the Food and Drug Administration (FDA)-authorized Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Targets (MSK-IMPACT) assay. Seven homologous recombination deficiency (HRD), two DNA damage response (DDR) and four mismatch repair (MMR) genes, as well as NF1, RB1 and TP53 were included in the analysis. FDA-authorized and New York State Department of Health-approved sequencing methods for germline, tumor/normal and tumor-only sequencing assays and analytical pipelines were employed.ResultsIn patients who underwent tumor and germline sequencing, as compared to clinical genetic testing, tumor-only sequencing failed to detect 10.5% of clinically actionable pathogenic germline variants in CSGs, including 18.8%, 12.8% and 7.3% of germline variants in MMR, DDR and HRD genes, respectively. The sensitivity for detection of pathogenic germline variants by tumor-only sequencing was 89.5%. Whilst the vast majority of pathogenic germline exonic single-nucleotide variants (SNVs) and small indels were detected by tumor-only sequencing, large percentages of germline copy number variants, intronic variants and repetitive element insertions were not detected.ConclusionsTumor-only sequencing is adequate for the detection of clinically actionable germline variants, particularly for SNVs and small indels; however, a small subset of alterations affecting HRD, DDR and MMR genes may not be detected optimally. Therefore, for high-risk patients with negative tumor-only sequencing results, clinical genetic testing could be considered given the impact of these variants on therapy and genetic counseling.