Accurately Identifying Low‐Allelic Fraction Variants in Single Samples with Next‐Generation Sequencing: Applications in Tumor Subclone Resolution

Current methods for resolving genetically distinct subclones in tumor samples require somatic mutations to be clustered by allelic frequencies, which are determined by applying a variant calling program to next‐generation sequencing data. Such programs were developed to accurately distinguish true polymorphisms and somatic mutations from the artifactual nonreference alleles introduced during library preparation and sequencing. However, numerous variant callers exist with no clear indication of the best performer for subclonal analysis, in which the accuracy of the assigned variant frequency is as important as correctly indicating whether the variant is present or not. Furthermore, sequencing depth (the number of times that a genomic position is sequenced) affects the ability to detect low‐allelic fraction variants and accurately assign their allele frequencies. We created two synthetic sequencing datasets, and sequenced real KRAS amplicons, with variants spiked in at specific ratios, to assess which caller performs best in terms of both variant detection and assignment of allelic frequencies. We also assessed the sequencing depths required to detect low‐allelic fraction variants. We found that VarScan2 performed best overall with sequencing depths of 100×, 250×, 500×, and 1,000× required to accurately identify variants present at 10%, 5%, 2.5%, and 1%, respectively.     

A Robust Approach for Blind Detection of Balanced Chromosomal Rearrangements with Whole‐Genome Low‐Coverage Sequencing

Balanced chromosomal rearrangement (or balanced chromosome abnormality, BCA) is a common chromosomal structural variation. Next‐generation sequencing has been reported to detect BCA‐associated breakpoints with the aid of karyotyping. However, the complications associated with this approach and the requirement for cytogenetics information has limited its application. Here, we provide a whole‐genome low‐coverage sequencing approach to detect BCA events independent of knowing the affected regions and with low false positives. First, six samples containing BCAs were used to establish a detection protocol and assess the efficacy of different library construction approaches. By clustering anomalous read pairs and filtering out the false‐positive results with a control cohort and the concomitant mapping information, we could directly detect BCA events for each sample. Through optimizing the read depth, BCAs in all samples could be blindly detected with only 120 million read pairs per sample for data from a small‐insert library and 30 million per sample for data from nonsize‐selected mate‐pair library. This approach was further validated using another 13 samples that contained BCAs. Our approach advances the application of high‐throughput whole‐genome low‐coverage analysis for robust BCA detection—especially for clinical samples—without the need for karyotyping.     

Amplicon Resequencing Identified Parental Mosaicism for Approximately 10% of “de novo” SCN1A Mutations in Children with Dravet Syndrome

The majority of children with Dravet syndrome (DS) are caused by de novo SCN1A mutations. To investigate the origin of the mutations, we developed and applied a new method that combined deep amplicon resequencing with a Bayesian model to detect and quantify allelic fractions with improved sensitivity. Of 174 SCN1A mutations in DS probands which were considered “de novo” by Sanger sequencing, we identified 15 cases (8.6%) of parental mosaicism. We identified another five cases of parental mosaicism that were also detectable by Sanger sequencing. Fraction of mutant alleles in the 20 cases of parental mosaicism ranged from 1.1% to 32.6%. Thirteen (65% of 20) mutations originated paternally and seven (35% of 20) maternally. Twelve (60% of 20) mosaic parents did not have any epileptic symptoms. Their mutant allelic fractions were significantly lower than those in mosaic parents with epileptic symptoms (P = 0.016). We identified mosaicism with varied allelic fractions in blood, saliva, urine, hair follicle, oral epithelium, and semen, demonstrating that postzygotic mutations could affect multiple somatic cells as well as germ cells. Our results suggest that more sensitive tools for detecting low‐level mosaicism in parents of families with seemingly “de novo” mutations will allow for better informed genetic counseling.     

Targeted Next‐Generation Sequencing can Replace Sanger Sequencing in Clinical Diagnostics

Mutation detection through exome sequencing allows simultaneous analysis of all coding sequences of genes. However, it cannot yet replace Sanger sequencing (SS) in diagnostics because of incomplete representation and coverage of exons leading to missing clinically relevant mutations. Targeted next‐generation sequencing (NGS), in which a selected fraction of genes is sequenced, may circumvent these shortcomings. We aimed to determine whether the sensitivity and specificity of targeted NGS is equal to those of SS. We constructed a targeted enrichment kit that includes 48 genes associated with hereditary cardiomyopathies. In total, 84 individuals with cardiomyopathies were sequenced using 151 bp paired‐end reads on an Illumina MiSeq sequencer. The reproducibility was tested by repeating the entire procedure for five patients. The coverage of ≥30 reads per nucleotide, our major quality criterion, was 99% and in total ～21,000 variants were identified. Confirmation with SS was performed for 168 variants (155 substitutions, 13 indels). All were confirmed, including a deletion of 18 bp and an insertion of 6 bp. The reproducibility was nearly 100%. We demonstrate that targeted NGS of a disease‐specific subset of genes is equal to the quality of SS and it can therefore be reliably implemented as a stand‐alone diagnostic test.     

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.     

Sensitive multistep clinical molecular screening of 180 unrelated individuals with retinoblastoma detects 36 novel mutations in the RB1 gene

Thymopoietin or TMPO (indicated by its alternative gene symbol, LAP2, in this work) has been proposed as a candidate disease gene for dilated cardiomyopathy (DCM), since a LAP2 product associates with nucleoplasmic lamins A/C, which are encoded by the DCM gene LMNA. We developed a study to screen for genetic mutations in LAP2 in a large collection of DCM patients and families. A total of 113 subjects from 88 families (56 with familial DCM (FDC) and 32 with sporadic DCM) were screened for LAP2 mutations using denaturing high-performance liquid chromatography and sequence analysis. We found a single putative mutation affecting the LAP2alpha isoform in one FDC pedigree. The mutation predicts an Arg690Cys substitution (c.2068C>T; p.R690C) located in the C-terminal domain of the LAP2alpha protein, a region that is known to interact with lamin A/C. RT-PCR, Western blot analyses, and immunolocalization revealed low-level LAP2alpha expression in adult cardiac muscle, and localization to a subset of nuclei. Mutated Arg690Cys LAP2alpha expressed in HeLa cells localized to the nucleoplasm like wild-type LAP2alpha, with no effect on peripheral and nucleoplasmic lamin A distribution. However, the in vitro interaction of mutated LAP2alpha with the pre-lamin A C-terminus was significantly compromised compared to the wild-type protein. LAP2 mutations may represent a rare cause of DCM. The Arg690Cys mutation altered the observed LAP2alpha interaction with A-type lamins. Our finding implicates a novel nuclear lamina-associated protein in the pathogenesis of genetic forms of dilated cardiomyopathy.     

Detection of mosaic RB1 mutations in families with retinoblastoma

The RB1 gene mutation detection rate in 1,020 retinoblastoma families was increased by the use of highly sensitive allele specific‐PCR (AS‐PCR) to detect low‐level mosaicism for 11 recurrent RB1 CGA>TGA nonsense mutations. For bilaterally affected probands, AS‐PCR increased the RB1 mutation detection sensitivity from 92.6% to 94.8%. Both RB1 oncogenic changes were detected in 92.7% of sporadic unilateral tumors (357/385); 14.6% (52/357) of unilateral probands with both tumor mutations identified carried one of the tumor mutations in blood. Mosaicism was evident in 5.5% of bilateral probands (23 of 421), in 3.8% of unilateral probands (22 of 572), and in one unaffected mother of a unilateral proband. Half of the mosaic mutations were only detectable by AS‐PCR for the 11 recurrent CGA>TGA mutations, and not by standard sequencing. This suggests that significant numbers of low‐level mosaics with other classes of RB1 mutations remain unidentified by current technology. We show that the use of linkage analysis in a two‐generation retinoblastoma family resulted in the erroneous conclusion that a child carried the parental mutation, because the founder parent was mosaic for the RB1 mutation. Of 142 unaffected parental pairs tested, only one unaffected parent of a proband (0.7%) showed somatic mosaicism for the proband's mutation, in contrast to an overall 4.5% somatic mosaicism rate for retinoblastoma probands, suggesting that mosaicism for an RB1 mutation is highly likely to manifest as retinoblastoma. Hum Mutat 0, 1–10, 2009. © 2009 Wiley‐Liss, Inc.     

Simultaneous Analysis of Hundreds of Y‐Chromosomal SNPs for High‐Resolution Paternal Lineage Classification using Targeted Semiconductor Sequencing

SNPs from the non‐recombining part of the human Y chromosome (Y‐SNPs) are informative to classify paternal lineages in forensic, genealogical, anthropological, and evolutionary studies. Although thousands of Y‐SNPs were identified thus far, previous Y‐SNP multiplex tools target only dozens of markers simultaneously, thereby restricting the provided Y‐haplogroup resolution and limiting their applications. Here, we overcome this shortcoming by introducing a high‐resolution multiplex tool for parallel genotyping‐by‐sequencing of 530 Y‐SNPs using the Ion Torrent PGM platform, which allows classification of 432 worldwide Y haplogroups. Contrary to previous Y‐SNP multiplex tools, our approach covers branches of the entire Y tree, thereby maximizing the paternal lineage classification obtainable. We used a default DNA input amount of 10 ng per reaction but preliminary sensitivity testing revealed positive results from as little as 100 pg input DNA. Furthermore, we demonstrate that sample pooling using barcodes is feasible, allowing increased throughput for lower per‐sample costs. In addition to the wetlab protocol, we provide a software tool for automated data quality control and haplogroup classification. The unique combination of ultra‐high marker density and high sensitivity achievable from low amounts of potentially degraded DNA makes this new multiplex tool suitable for a wide range of Y‐chromosome applications.     

A Short‐Read Multiplex Sequencing Method for Reliable,Cost‐Effective and High‐Throughput Genotyping in Large‐Scale Studies

Accurate genotyping is important for genetic testing. Sanger sequencing‐based typing is the gold standard for genotyping, but it has been underused, due to its high cost and low throughput. In contrast, short‐read sequencing provides inexpensive and high‐throughput sequencing, holding great promise for reaching the goal of cost‐effective and high‐throughput genotyping. However, the short‐read length and the paucity of appropriate genotyping methods, pose a major challenge. Here, we present RCHSBT—reliable, cost‐effective and high‐throughput sequence based typing pipeline—which takes short sequence reads as input, but uses a unique variant calling, haploid sequence assembling algorithm, can accurately genotype with greater effective length per amplicon than even Sanger sequencing reads. The RCHSBT method was tested for the human MHC loci HLA‐A, HLA‐B, HLA‐C, HLA‐DQB1, and HLA‐DRB1, upon 96 samples using Illumina PE 150 reads. Amplicons as long as 950 bp were readily genotyped, achieving 100% typing concordance between RCHSBT‐called genotypes and genotypes previously called by Sanger sequence. Genotyping throughput was increased over 10 times, and cost was reduced over five times, for RCHSBT as compared with Sanger sequence genotyping. We thus demonstrate RCHSBT to be a genotyping method comparable to Sanger sequencing‐based typing in quality, while being more cost‐effective, and higher throughput.     

Integrated Sequence Analysis Pipeline Provides One‐Stop Solution for Identifying Disease‐Causing Mutations

Next‐generation sequencing has greatly accelerated the search for disease‐causing defects, but even for experts the data analysis can be a major challenge. To facilitate the data processing in a clinical setting, we have developed a novel medical resequencing analysis pipeline (MERAP). MERAP assesses the quality of sequencing, and has optimized capacity for calling variants, including single‐nucleotide variants, insertions and deletions, copy‐number variation, and other structural variants. MERAP identifies polymorphic and known causal variants by filtering against public domain databases, and flags nonsynonymous and splice‐site changes. MERAP uses a logistic model to estimate the causal likelihood of a given missense variant. MERAP considers the relevant information such as phenotype and interaction with known disease‐causing genes. MERAP compares favorably with GATK, one of the widely used tools, because of its higher sensitivity for detecting indels, its easy installation, and its economical use of computational resources. Upon testing more than 1,200 individuals with mutations in known and novel disease genes, MERAP proved highly reliable, as illustrated here for five families with disease‐causing variants. We believe that the clinical implementation of MERAP will expedite the diagnostic process of many disease‐causing defects.     

ICO Amplicon NGS Data Analysis: A Web Tool for Variant Detection in Common High‐Risk Hereditary Cancer Genes Analyzed by Amplicon GS Junior Next‐Generation Sequencing

Next‐generation sequencing (NGS) has revolutionized genomic research and is set to have a major impact on genetic diagnostics thanks to the advent of benchtop sequencers and flexible kits for targeted libraries. Among the main hurdles in NGS are the difficulty of performing bioinformatic analysis of the huge volume of data generated and the high number of false positive calls that could be obtained, depending on the NGS technology and the analysis pipeline. Here, we present the development of a free and user‐friendly Web data analysis tool that detects and filters sequence variants, provides coverage information, and allows the user to customize some basic parameters. The tool has been developed to provide accurate genetic analysis of targeted sequencing of common high‐risk hereditary cancer genes using amplicon libraries run in a GS Junior System. The Web resource is linked to our own mutation database, to assist in the clinical classification of identified variants. We believe that this tool will greatly facilitate the use of the NGS approach in routine laboratories.     

mit‐o‐matic: A Comprehensive Computational Pipeline for Clinical Evaluation of Mitochondrial Variations from Next‐Generation Sequencing Datasets

The human mitochondrial genome has been reported to have a very high mutation rate as compared with the nuclear genome. A large number of mitochondrial mutations show significant phenotypic association and are involved in a broad spectrum of diseases. In recent years, there has been a remarkable progress in the understanding of mitochondrial genetics. The availability of next‐generation sequencing (NGS) technologies have not only reduced sequencing cost by orders of magnitude but has also provided us good quality mitochondrial genome sequences with high coverage, thereby enabling decoding of a number of human mitochondrial diseases. In this study, we report a computational and experimental pipeline to decipher the human mitochondrial DNA variations and examine them for their clinical correlation. As a proof of principle, we also present a clinical study of a patient with Leigh disease and confirmed maternal inheritance of the causative allele. The pipeline is made available as a user‐friendly online tool to annotate variants and find haplogroup, disease association, and heteroplasmic sites. The “mit‐o‐matic” computational pipeline represents a comprehensive cloud‐based tool for clinical evaluation of mitochondrial genomic variations from NGS datasets. The tool is freely available at http://genome.igib.res.in/mitomatic/ .     

Mutation Detection by Clonal Sequencing of PCR Amplicons and Grouped Read Typing is Applicable to Clinical Diagnostics

We describe a sensitive technique for mutation detection using clonal sequencing. We analyzed DNA extracted from 13 cancer cell lines and 35 tumor samples and applied a novel approach to identify disease‐associated somatic mutations. By matching reads against an index of known variants, noise can be dramatically reduced, enabling the detection and quantification of those variants, even when they are present at less than 1% of the total sequenced population; this is comparable to, or better than, current diagnostic methods. Following the identification or exclusion of known variants, unmatched reads are grouped for BLAST searching to identify novel variants or contaminants. Known variants, novel variants, and contaminants were readily identified in tumor tissue using this approach. Our approach also enables an estimation of the per‐base sequencing error rate, providing a confidence threshold for interpretation of the results in the clinic. This novel approach has immediate applicability to clinical testing for disease‐associated genetic variants.     

Evaluation of Hybridization Capture Versus Amplicon‐Based Methods for Whole‐Exome Sequencing

Next‐generation sequencing has aided characterization of genomic variation. While whole‐genome sequencing may capture all possible mutations, whole‐exome sequencing remains cost‐effective and captures most phenotype‐altering mutations. Initial strategies for exome enrichment utilized a hybridization‐based capture approach. Recently, amplicon‐based methods were designed to simplify preparation and utilize smaller DNA inputs. We evaluated two hybridization capture‐based and two amplicon‐based whole‐exome sequencing approaches, utilizing both Illumina and Ion Torrent sequencers, comparing on‐target alignment, uniformity, and variant calling. While the amplicon methods had higher on‐target rates, the hybridization capture‐based approaches demonstrated better uniformity. All methods identified many of the same single‐nucleotide variants, but each amplicon‐based method missed variants detected by the other three methods and reported additional variants discordant with all three other technologies. Many of these potential false positives or negatives appear to result from limited coverage, low variant frequency, vicinity to read starts/ends, or the need for platform‐specific variant calling algorithms. All methods demonstrated effective copy‐number variant calling when evaluated against a single‐nucleotide polymorphism array. This study illustrates some differences between whole‐exome sequencing approaches, highlights the need for selecting appropriate variant calling based on capture method, and will aid laboratories in selecting their preferred approach.     

Analysis of Human Triallelic SNPs by Next‐Generation Sequencing

Although single‐nucleotide polymorphisms (SNPs) have become extremely useful in the study of geneticvariation, triallelic SNPs are still not fully understood. Next‐generation sequencing (NGS) is a promising approach to identify triallelic sites in large populations. In this study, we explored exome sequencing data from 221 Chinese individuals, with an average depth of 70‐fold. We identified 382,901 SNPs in the study samples, including 2,002 (0.52%) triallelic sites. Among the triallelic SNPs, 17.3% were coding SNPs (cSNPs) and 78.3% were novel. Comparison and analysis revealed that the variant alleles were more likely to result in nonsynonymous variation at triallelic sites. In addition, natural selection seemed to influence triallelic SNPs. However, with the limited sample size assessed, more studies will be required in order to fully characterize the features of triallelic SNPs.     

Next‐generation DNA sequencing of a Swedish malignant hyperthermia cohort

Malignant hyperthermia (MH)‐related mutations have been identified in the ryanodine receptor type 1 gene (RYR1) and in the dihydropyridine gene (CACNA1S), but about half of the patients do not have causative mutations in these genes. We wanted to study the contribution of other muscle genes to the RYR1 phenotypes. We designed a gene panel for sequence enrichment targeting 64 genes of proteins involved in the homeostasis of the striated muscle cell. Next‐generation sequencing (NGS) resulted in >50,000 sequence variants which were further analyzed by software filtering criteria to identify causative variants. In four of five patients we identified previously reported RYR1 mutations while the fifth patient did not show any candidate variant in any of the genes investigated. In two patients pathogenic variants were found in other genes known to cause a muscle disorders. All but one patient carried likely benign rare polymorphisms. The NGS technique proved convenient in identifying variants in the RYR1. However, with a clinically variable phenotype‐like MH, the pre‐selection of genes poses problems in variant interpretation.     

Targeted and Genomewide NGS Data Disqualify Mutations in MYO1A,the “DFNA48 Gene”, as a Cause of Deafness

MYO1A is considered the gene underlying autosomal dominant nonsyndromic hearing loss DFNA48, based on six missense variants, one small in‐frame insertion, and one nonsense mutation. Results from NGS targeting 66 deafness genes in 109 patients identified three families challenging this assumption: two novel nonsense (p.Tyr740* and p.Arg262*) and a known missense variant were identified heterozygously not only in index patients, but also in unaffected relatives. Deafness in these families clearly resulted from mutations in other genes (MYO7A, EYA1, and CIB2). Most of the altogether 10 MYO1A mutations are annotated in dbSNP, and population frequencies (dbSNP, 1000 Genomes, Exome Sequencing Project) above 0.1% contradict pathogenicity under a dominant model. One healthy individual was even homozygous for p.Arg262*, compatible with homozygous Myo1a knockout mice lacking any overt pathology. MYO1A seems dispensable for hearing and overall nonessential. MYO1A adds to the list of “erroneous disease genes”, which will expand with increasing availability of large‐scale sequencing data.     

Targeted Next‐Generation Sequencing Analysis of 1,000 Individuals with Intellectual Disability

To identify genetic causes of intellectual disability (ID), we screened a cohort of 986 individuals with moderate to severe ID for variants in 565 known or candidate ID‐associated genes using targeted next‐generation sequencing. Likely pathogenic rare variants were found in ～11% of the cases (113 variants in 107/986 individuals: ～8% of the individuals had a likely pathogenic loss‐of‐function [LoF] variant, whereas ～3% had a known pathogenic missense variant). Variants in SETD5, ATRX, CUL4B, MECP2, and ARID1B were the most common causes of ID. This study assessed the value of sequencing a cohort of probands to provide a molecular diagnosis of ID, without the availability of DNA from both parents for de novo sequence analysis. This modeling is clinically relevant as 28% of all UK families with dependent children are single parent households. In conclusion, to diagnose patients with ID in the absence of parental DNA, we recommend investigation of all LoF variants in known genes that cause ID and assessment of a limited list of proven pathogenic missense variants in these genes. This will provide 11% additional diagnostic yield beyond the 10%–15% yield from array CGH alone.     

New Tools for Mendelian Disease Gene Identification: PhenoDB Variant Analysis Module; and GeneMatcher,a Web‐Based Tool for Linking Investigators with an Interest in the Same Gene

Identifying the causative variant from among the thousands identified by whole‐exome sequencing or whole‐genome sequencing is a formidable challenge. To make this process as efficient and flexible as possible, we have developed a Variant Analysis Module coupled to our previously described Web‐based phenotype intake tool, PhenoDB ( http://researchphenodb.net and http://phenodb.org ). When a small number of candidate‐causative variants have been identified in a study of a particular patient or family, a second, more difficult challenge becomes proof of causality for any given variant. One approach to this problem is to find other cases with a similar phenotype and mutations in the same candidate gene. Alternatively, it may be possible to develop biological evidence for causality, an approach that is assisted by making connections to basic scientists studying the gene of interest, often in the setting of a model organism. Both of these strategies benefit from an open access, online site where individual clinicians and investigators could post genes of interest. To this end, we developed GeneMatcher ( http://genematcher.org ), a freely accessible Website that enables connections between clinicians and researchers across the world who share an interest in the same gene(s).     

Lake Louise Mutation Detection Meeting 2013: Clinical Translation of Next‐Generation Sequencing Requires Optimization of Workflows and Interpretation of Variants

With the exponential reduction of the cost of next‐generation sequencing (NGS), it is no longer the generation of data but the analysis and interpretation of massive amounts of sequencing data that are seen as key challenges for the effective integration of these technologies into clinical practice. Clinical geneticists, informaticians, and scientists from 17 countries gathered for the 12th International Symposium on Mutation in the Genome at the Fairmont Chateau Lake Louise (Canada) to discuss technological advances and applications of NGS and consider possible approaches to the challenges of clinical translation. Here, we provide an overview of the main themes of the meeting that included development of innovative solutions for variant sharing, tools and resources for NGS analysis, novel technology and methodology development, NGS‐based discovery of disease pathogenesis, development of multigene NGS sequencing panels for clinical use, exploring diagnostic utility of whole‐exome and whole‐genome sequencing, and, finally, integration of genomic sequencing into the clinic.