Mutascope: sensitive detection of somatic mutations from deep amplicon sequencing

LoLoPicker: detecting low allelic-fraction variants from low-quality cancer samples

Introduction

Although several programs are designed to identify variants with low allelic-fraction, further improvement is needed, especially to push the detection limit of low allelic-faction variants in low-quality, ”noisy” tumor samples.

Results

We developed LoLoPicker, an efficient tool dedicated to calling somatic variants from next-generation sequencing (NGS) data of tumor sample against the matched normal sample plus a user-defined control panel of additional normal samples. The control panel allows accurately estimating background error rate and therefore ensures high-accuracy mutation detection.

Conclusions

Compared to other methods, we showed a superior performance of LoLoPicker with significantly improved specificity. The algorithm of LoLoPicker is particularly useful for calling low allelic-fraction variants from low-quality cancer samples such as formalin-fixed and paraffin-embedded (FFPE) samples.

Implementation and Availability: The main scripts are implemented in Python-2.7 and the package is released athttps://github.com/jcarrotzhang/LoLoPicker.

Assessment of Capture and Amplicon-Based Approaches for the Development of a Targeted Next-Generation Sequencing Pipeline to Personalize Lymphoma Management

Targeted next-generation sequencing panels are increasingly used to assess the value of gene mutations for clinical diagnostic purposes. For assay development, amplicon-based methods have been preferentially used on the basis of short preparation time and small DNA input amounts. However, capture sequencing has emerged as an alternative approach because of high testing accuracy. We compared capture hybridization and amplicon sequencing approaches using fresh-frozen and formalin-fixed, paraffin-embedded tumor samples from eight lymphoma patients. Next, we developed a targeted sequencing pipeline using a 32-gene panel for accurate detection of actionable mutations in formalin-fixed, paraffin-embedded tumor samples of the most common lymphocytic malignancies: chronic lymphocytic leukemia, diffuse large B-cell lymphoma, and follicular lymphoma. We show that hybrid capture is superior to amplicon sequencing by providing deep more uniform coverage and yielding higher sensitivity for variant calling. Sanger sequencing of 588 variants identified specificity limits of thresholds for mutation calling, and orthogonal validation on 66 cases indicated 93% concordance with whole-genome sequencing. The developed pipeline and assay identified at least one actionable mutation in 91% of tumors from 219 lymphoma patients and revealed subtype-specific mutation patterns and frequencies consistent with the literature. This pipeline is an accurate and sensitive method for identifying actionable gene mutations in routinely acquired biopsy materials, suggesting further assessment of capture-based assays in the context of personalized lymphoma management.

Canary: an atomic pipeline for clinical amplicon assays

BACKGROUND

High throughput sequencing requires bioinformatics pipelines to process large volumes of data into meaningful variants that can be translated into a clinical report. These pipelines often suffer from a number of shortcomings: they lack robustness and have many components written in multiple languages, each with a variety of resource requirements. Pipeline components must be linked together with a workflow system to achieve the processing of FASTQ files through to a VCF file of variants. Crafting these pipelines requires considerable bioinformatics and IT skills beyond the reach of many clinical laboratories.

RESULTS

Here we present Canary, a single program that can be run on a laptop, which takes FASTQ files from amplicon assays through to an annotated VCF file ready for clinical analysis. Canary can be installed and run with a single command using Docker containerization or run as a single JAR file on a wide range of platforms. Although it is a single utility, Canary performs all the functions present in more complex and unwieldy pipelines. All variants identified by Canary are 3' shifted and represented in their most parsimonious form to provide a consistent nomenclature, irrespective of sequencing variation. Further, proximate in-phase variants are represented as a single HGVS 'delins' variant. This allows for correct nomenclature and consequences to be ascribed to complex multi-nucleotide polymorphisms (MNPs), which are otherwise difficult to represent and interpret. Variants can also be annotated with hundreds of attributes sourced from MyVariant.info to give up to date details on pathogenicity, population statistics and in-silico predictors.

CONCLUSIONS

Canary has been used at the Peter MacCallum Cancer Centre in Melbourne for the last 2 years for the processing of clinical sequencing data. By encapsulating clinical features in a single, easily installed executable, Canary makes sequencing more accessible to all pathology laboratories. Canary is available for download as source or a Docker image at https://github.com/PapenfussLab/Canary under a GPL-3.0 License.

Identifying DNase I hypersensitive sites as driver distal regulatory elements in breast cancer

Efforts to identify driver mutations in cancer have largely focused on genes, whereas non-coding sequences remain relatively unexplored. Here we develop a statistical method based on characteristics known to influence local mutation rate and a series of enrichment filters in order to identify distal regulatory elements harboring putative driver mutations in breast cancer. We identify ten DNase I hypersensitive sites that are significantly mutated in breast cancers and associated with the aberrant expression of neighboring genes. A pan-cancer analysis shows that three of these elements are significantly mutated across multiple cancer types and have mutation densities similar to protein-coding driver genes. Functional characterization of the most highly mutated DNase I hypersensitive sites in breast cancer (using in silico and experimental approaches) confirms that they are regulatory elements and affect the expression of cancer genes. Our study suggests that mutations of regulatory elements in tumors likely play an important role in cancer development.

Cancer driver mutations can occur within noncoding genomic sequences. Here, the authors develop a statistical approach to identify candidate noncoding driver mutations in DNase I hypersensitive sites in breast cancer and experimentally demonstrate they are regulatory elements of known cancer genes.

Ultrasensitive detection of acute myeloid leukemia minimal residual disease using single molecule molecular inversion probes

The identification of minimal residual disease is the primary diagnostic finding which predicts relapse in patients treated for acute myeloid leukemia. Ultrasensitive detection of minimal residual disease would enable better patient risk stratification and could open opportunities for early therapeutic intervention. Herein we apply single molecule molecular inversion probe capture, a technology combining multiplexed targeted sequencing with error correction schemes based on molecular barcoding, in order to detect mutations identifying minimal residual disease with ultrasensitive and quantitative precision. We designed a single molecule molecular inversion probe capture panel spanning >50 kb and targeting 32 factors relevant to acute myeloid leukemia pathogenesis. We demonstrate linearity and quantitative precision over 100-fold relative abundance of mutant cells (1 in 100 to 1 in 1,500), with estimated error rates approaching 1 in 1,200 base pairs sequenced and maximum theoretical limits of detection exceeding 1 in 60,000 mutant alleles. In 3 of 4 longitudinally collected specimens from patients with acute myeloid leukemia, we find that single molecule molecular inversion probe capture detects somatic mutations identifying minimal residual disease at substantially earlier time points and with greater sensitivity than clinical diagnostic approaches used as current standard of care (flow cytometry and conventional molecular diagnosis), and identifies persisting neoplastic cells during clinical remission. In 2 patients, single molecule molecular inversion probe capture detected heterogeneous, subclonal acute myeloid leukemia populations carrying distinct mutational signatures. Single molecule molecular inversion probe technology uniquely couples scalable target enrichment with sequence read error correction, providing an integrated, ultrasensitive approach for detecting minimal residual disease identifying mutations.

Variant profiling of evolving prokaryotic populations

Genomic heterogeneity of bacterial species is observed and studied in experimental evolution experiments and clinical diagnostics, and occurs as micro-diversity of natural habitats. The challenge for genome research is to accurately capture this heterogeneity with the currently used short sequencing reads. Recent advances in NGS technologies improved the speed and coverage and thus allowed for deep sequencing of bacterial populations. This facilitates the quantitative assessment of genomic heterogeneity, including low frequency alleles or haplotypes. However, false positive variant predictions due to sequencing errors and mapping artifacts of short reads need to be prevented. We therefore created VarCap, a workflow for the reliable prediction of different types of variants even at low frequencies. In order to predict SNPs, InDels and structural variations, we evaluated the sensitivity and accuracy of different software tools using synthetic read data. The results suggested that the best sensitivity could be reached by a union of different tools, however at the price of increased false positives. We identified possible reasons for false predictions and used this knowledge to improve the accuracy by post-filtering the predicted variants according to properties such as frequency, coverage, genomic environment/localization and co-localization with other variants. We observed that best precision was achieved by using an intersection of at least two tools per variant. This resulted in the reliable prediction of variants above a minimum relative abundance of 2%. VarCap is designed for being routinely used within experimental evolution experiments or for clinical diagnostics. The detected variants are reported as frequencies within a VCF file and as a graphical overview of the distribution of the different variant/allele/haplotype frequencies. The source code of VarCap is available at https://github.com/ma2o/VarCap. In order to provide this workflow to a broad community, we implemeted VarCap on a Galaxy webserver, which is accessible at http://galaxy.csb.univie.ac.at.

Analysis of amplicon-based NGS data from neurological disease gene panels: a new method for allele drop-out management

BACKGROUND

Amplicon-based targeted resequencing is a commonly adopted solution for next-generation sequencing applications focused on specific genomic regions. The reliability of such approaches rests on the high specificity and deep coverage, although sequencing artifacts attributable to PCR-like amplification can be encountered. Between these artifacts, allele drop-out, which is the preferential amplification of one allele, causes an artificial increase in homozygosity when heterozygous mutations fall on a primer pairing region. Here, a procedure to manage such artifacts, based on a pipeline composed of two steps of alignment and variant calling, is proposed. This methodology has been compared to the Illumina Custom Amplicon workflow, available on Illumina MiSeq, on the analysis of data obtained with four newly designed TruSeq Custom Amplicon gene panels.

RESULTS

Four gene panels, specific for Parkinson disease, for Intracerebral Hemorrhage Diseases (COL4A1 and COL4A2 genes) and for Familial Hemiplegic Migraine (CACNA1A and ATP1A2 genes) were designed. A total of 119 samples were re-sequenced with Illumina MiSeq sequencer and panel characterization in terms of coverage, number of variants found and allele drop-out potential impact has been carried out. Results show that 14 % of identified variants is potentially affected by allele drop-out artifacts and that both the Custom Amplicon workflow and the procedure proposed here could correctly identify them. Furthermore, a more complex configuration in presence of two mutations was simulated in silico. In this configuration, our proposed methodology outperforms Custom Amplicon workflow, being able to correctly identify two mutations in all the studied configurations.

CONCLUSIONS

Allele drop-out plays a crucial role in amplicon-based targeted re-sequencing and specific procedures in data analysis of amplicon data should be adopted. Although a consensus has been established in the elimination of primer sequences from aligned data (e.g., via primer sequence trimming or soft clipping), more complex configurations need to be managed in order to increase the retrieved information from available data. Our method shows how to manage one of these complex configurations, when two mutations occur.

Recurrent activating mutations of CD28 in peripheral T-cell lymphomas

Peripheral T-cell lymphomas (PTCLs) comprise a heterogeneous group of mature T-cell neoplasms with a poor prognosis. Recently, mutations in TET2 and other epigenetic modifiers as well as RHOA have been identified in these diseases, particularly in angioimmunoblastic T-cell lymphoma (AITL). CD28 is the major co-stimulatory receptor in T cells which, upon binding ligand, induces sustained T-cell proliferation and cytokine production when combined with T-cell receptor stimulation. We have identified recurrent mutations in CD28 in PTCLs. Two residues-D124 and T195-were recurrently mutated in 11.3% of cases of AITL and in one case of PTCL, not otherwise specified (PTCL-NOS). Surface plasmon resonance analysis of mutations at these residues with predicted differential partner interactions showed increased affinity for ligand CD86 (residue D124) and increased affinity for intracellular adaptor proteins GRB2 and GADS/GRAP2 (residue T195). Molecular modeling studies on each of these mutations suggested how these mutants result in increased affinities. We found increased transcription of the CD28-responsive genes CD226 and TNFA in cells expressing the T195P mutant in response to CD3 and CD86 co-stimulation and increased downstream activation of NF-κB by both D124V and T195P mutants, suggesting a potential therapeutic target in CD28-mutated PTCLs.

Fast and scalable inference of multi-sample cancer lineages

Somatic variants can be used as lineage markers for the phylogenetic reconstruction of cancer evolution. Since somatic phylogenetics is complicated by sample heterogeneity, novel specialized tree-building methods are required for cancer phylogeny reconstruction. We present LICHeE (Lineage Inference for Cancer Heterogeneity and Evolution), a novel method that automates the phylogenetic inference of cancer progression from multiple somatic samples. LICHeE uses variant allele frequencies of somatic single nucleotide variants obtained by deep sequencing to reconstruct multi-sample cell lineage trees and infer the subclonal composition of the samples. LICHeE is open source and available at http://viq854.github.io/lichee.

BAP1 mutation is a frequent somatic event in peritoneal malignant mesothelioma

BACKGROUND

Malignant mesothelioma (MM) arises from mesothelial cells that line the pleural, peritoneal and pericardial surfaces. The majority of MMs are pleural and have been associated with asbestos exposure. Previously, pleural MMs have been genetically characterized by the loss of BAP1 (40-60%) as well as loss of NF2 (75%) and CDKN2A (60%). The rare peritoneal form of MM occurs in ~10% cases. With only ~300 cases diagnosed in the US per year, its link to asbestos exposure is not clear and its mutational landscape unknown.

METHODS

We analyzed the somatic mutational landscape of 12 peritoneal MM of epitheloid subtype using copy number analysis (N = 9), whole exome sequencing (N = 7) and targeted sequencing (N = 12).

RESULTS

Peritoneal MM display few copy number alterations, with most samples having less than 10 Mbp total changes, mostly through deletions and no high copy number amplification. Chromosome band 3p21 encoding BAP1 is the most recurrently deleted region (5/9), while, in contrast to pleural MM, NF2 and CDKN2A are not affected. We further identified 87 non-silent mutations across 7 sequenced tumors, with a median of 8 mutated genes per tumor, resulting in a very low mutation rate (median 1.3 10(-6)). BAP1 was the only recurrently mutated gene (N = 3/7). In one additional case, loss of the entire chromosome 3 leaves a non-functional copy of BAP1 carrying a rare nonsense germline variant, thus suggesting a potential genetic predisposition in this patient. Finally, with targeted sequencing of BAP1 in 3 additional cases, we conclude that BAP1 is frequently altered through copy number losses (N = 3/12), mutations (N = 3/12) or both (N = 2/12) sometimes at a sub-clonal level.

CONCLUSION

Our findings suggest a major role for BAP1 in peritoneal MM susceptibility and oncogenesis and indicate important molecular differences to pleural MM as well as potential strategies for therapy and prevention.

AmpliVar: mutation detection in high-throughput sequence from amplicon-based libraries

Conventional means of identifying variants in high-throughput sequencing align each read against a reference sequence, and then call variants at each position. Here, we demonstrate an orthogonal means of identifying sequence variation by grouping the reads as amplicons prior to any alignment. We used AmpliVar to make key-value hashes of sequence reads and group reads as individual amplicons using a table of flanking sequences. Low-abundance reads were removed according to a selectable threshold, and reads above this threshold were aligned as groups, rather than as individual reads, permitting the use of sensitive alignment tools. We show that this approach is more sensitive, more specific, and more computationally efficient than comparable methods for the analysis of amplicon-based high-throughput sequencing data. The method can be extended to enable alignment-free confirmation of variants seen in hybridization capture target-enrichment data.

Tumor evolution and intratumor heterogeneity of an epithelial ovarian cancer investigated using next-generation sequencing

BACKGROUND

The extent to which metastatic tumors further evolve by accumulating additional mutations is unclear and has yet to be addressed extensively using next-generation sequencing of high-grade serous ovarian cancer.

METHODS

Eleven spatially separated tumor samples from the primary tumor and associated metastatic sites and two normal samples were obtained from a Stage IIIC ovarian cancer patient during cytoreductive surgery prior to chemotherapy. Whole exome sequencing and copy number analysis were performed. Omental exomes were sequenced with a high depth of coverage to thoroughly explore the variants in metastatic lesions. Somatic mutations were further validated by ultra-deep targeted sequencing to sort out false positives and false negatives. Based on the somatic mutations and copy number variation profiles, a phylogenetic tree was generated to explore the evolutionary relationship among tumor samples.

RESULTS

Only 6% of the somatic mutations were present in every sample of a given case with TP53 as the only known mutant gene consistently present in all samples. Two non-spatial clusters of primary tumors (cluster P1 and P2), and a cluster of metastatic regions (cluster M) were identified. The patterns of mutations indicate that cluster P1 and P2 diverged in the early phase of tumorigenesis, and that metastatic cluster M originated from the common ancestral clone of cluster P1 with few somatic mutations and copy number variations.

CONCLUSIONS

Although a high level of intratumor heterogeneity was evident in high-grade serous ovarian cancer, our results suggest that transcoelomic metastasis arises with little accumulation of somatic mutations and copy number alterations in this patient.

MADGiC: a model-based approach for identifying driver genes in cancer

Motivation: Identifying and prioritizing somatic mutations is an important and challenging area of cancer research that can provide new insights into gene function as well as new targets for drug development. Most methods for prioritizing mutations rely primarily on frequency-based criteria, where a gene is identified as having a driver mutation if it is altered in significantly more samples than expected according to a background model. Although useful, frequency-based methods are limited in that all mutations are treated equally. It is well known, however, that some mutations have no functional consequence, while others may have a major deleterious impact. The spatial pattern of mutations within a gene provides further insight into their functional consequence. Properly accounting for these factors improves both the power and accuracy of inference. Also important is an accurate background model.

Results: Here, we develop a Model-based Approach for identifying Driver Genes in Cancer (termed MADGiC) that incorporates both frequency and functional impact criteria and accommodates a number of factors to improve the background model. Simulation studies demonstrate advantages of the approach, including a substantial increase in power over competing methods. Further advantages are illustrated in an analysis of ovarian and lung cancer data from The Cancer Genome Atlas (TCGA) project.

Availability and implementation: R code to implement this method is available at http://www.biostat.wisc.edu/ kendzior/MADGiC/.

Contact: kendzior@biostat.wisc.edu

Supplementary information: Supplementary data are available at Bioinformatics online.

Postzygotic single-nucleotide mosaicisms in whole-genome sequences of clinically unremarkable individuals

Postzygotic single-nucleotide mutations (pSNMs) have been studied in cancer and a few other overgrowth human disorders at whole-genome scale and found to play critical roles. However, in clinically unremarkable individuals, pSNMs have never been identified at whole-genome scale largely due to technical difficulties and lack of matched control tissue samples, and thus the genome-wide characteristics of pSNMs remain unknown. We developed a new Bayesian-based mosaic genotyper and a series of effective error filters, using which we were able to identify 17 SNM sites from ~80× whole-genome sequencing of peripheral blood DNAs from three clinically unremarkable adults. The pSNMs were thoroughly validated using pyrosequencing, Sanger sequencing of individual cloned fragments, and multiplex ligation-dependent probe amplification. The mutant allele fraction ranged from 5%-31%. We found that C→T and C→A were the predominant types of postzygotic mutations, similar to the somatic mutation profile in tumor tissues. Simulation data showed that the overall mutation rate was an order of magnitude lower than that in cancer. We detected varied allele fractions of the pSNMs among multiple samples obtained from the same individuals, including blood, saliva, hair follicle, buccal mucosa, urine, and semen samples, indicating that pSNMs could affect multiple sources of somatic cells as well as germ cells. Two of the adults have children who were diagnosed with Dravet syndrome. We identified two non-synonymous pSNMs in SCN1A, a causal gene for Dravet syndrome, from these two unrelated adults and found that the mutant alleles were transmitted to their children, highlighting the clinical importance of detecting pSNMs in genetic counseling.

Quick, sensitive and specific detection and evaluation of quantification of minor variants by high-throughput sequencing

Minor variants have significant implications in quasispecies evolution, early cancer detection and non-invasive fetal genotyping but their accurate detection by next-generation sequencing (NGS) is hampered by sequencing errors. We generated sequencing data from mixtures at predetermined ratios in order to provide insight into sequencing errors and variations that can arise for which simulation cannot be performed. The information also enables better parameterization in depth of coverage, read quality and heterogeneity, library preparation techniques, technical repeatability for mathematical modeling, theory development and simulation experimental design. We devised minor variant authentication rules that achieved 100% accuracy in both testing and validation experiments. The rules are free from tedious inspection of alignment accuracy, sequencing read quality or errors introduced by homopolymers. The authentication processes only require minor variants to: (1) have minimum depth of coverage larger than 30; (2) be reported by (a) four or more variant callers, or (b) DiBayes or LoFreq, plus SNVer (or BWA when no results are returned by SNVer), and with the interassay coefficient of variation (CV) no larger than 0.1. Quantification accuracy undermined by sequencing errors could neither be overcome by ultra-deep sequencing, nor recruiting more variant callers to reach a consensus, such that consistent underestimation and overestimation (i.e. low CV) were observed. To accommodate stochastic error and adjust the observed ratio within a specified accuracy, we presented a proof of concept for the use of a double calibration curve for quantification, which provides an important reference towards potential industrial-scale fabrication of calibrants for NGS.

Evaluation of ultra-deep targeted sequencing for personalized breast cancer care

INTRODUCTION

The increasing number of targeted therapies, together with a deeper understanding of cancer genetics and drug response, have prompted major healthcare centers to implement personalized treatment approaches relying on high-throughput tumor DNA sequencing. However, the optimal way to implement this transformative methodology is not yet clear. Current assays may miss important clinical information such as the mutation allelic fraction, the presence of sub-clones or chromosomal rearrangements, or the distinction between inherited variants and somatic mutations. Here, we present the evaluation of ultra-deep targeted sequencing (UDT-Seq) to generate and interpret the molecular profile of 38 breast cancer patients from two academic medical centers.

METHODS

We sequenced 47 genes in matched germline and tumor DNA samples from 38 breast cancer patients. The selected genes, or the pathways they belong to, can be targeted by drugs or are important in familial cancer risk or drug metabolism.

RESULTS

Relying on the added value of sequencing matched tumor and germline DNA and using a dedicated analysis, UDT-Seq has a high sensitivity to identify mutations in tumors with low malignant cell content. Applying UDT-Seq to matched tumor and germline specimens from the 38 patients resulted in a proposal for at least one targeted therapy for 22 patients, the identification of tumor sub-clones in 3 patients, the suggestion of potential adverse drug effects in 3 patients and a recommendation for genetic counseling for 2 patients.

CONCLUSION

Overall our study highlights the additional benefits of a sequencing strategy, which includes germline DNA and is optimized for heterogeneous tumor tissues.