Introduction

Hereditary breast and ovarian cancer syndrome is caused by germline mutations in BRCA1 and BRCA2.1 Life time risks of breast cancer in carriers of these genes rise to 80%, and those of ovarian cancer are greater than 50 and 20% for carriers of BRCA1 and BRCA2, respectively.2 The identification of pathogenic mutations in BRCA1 and BRCA2 allows for prophylactic surgical intervention, which can markedly decrease cancer incidence, morbidity, and mortality.3, 4, 5, 6 The information about BRCA1/2 mutation status in a patient can also be useful to other family members. Genetic testing for BRCA1 and BRCA2 is conducted in many countries.7 Genetic testing for BRCA1 and BRCA2 mutations has become standard clinical practice for women with personal or family histories of breast or ovarian cancer. Genetic screening for BRCA1/2 mutations in the general population has been recently proposed, which may have potential benefit for early identification of carriers with pathogenic germline mutations.8 For women with pathogenic mutations, routine surveillance for breast cancer is recommended from age 25 and prophylactic surgical intervention is recommended after age 35 or when childbearing decisions are completed.7

Genetic testing for mutations in BRCA1 and BRCA2 has long been conducted with Sanger sequencing. According to the 2014 BRCA scheme report from the European Molecular Genetics Quality Network (EMQN), 75% of laboratories use Sanger sequencing alone for BRCA testing.9 Although Sanger sequencing is accurate and considered as ‘gold standard’ in clinical genetic testing, the major limitation is that this approach is laborious and time-consuming because BRCA1 and BRCA2 are large genes spanning approximately 81.1 and 84.7 kb and consisting of 24 and 27 exons, respectively. Additionally, Sanger sequencing requires a significant quantity of input DNA.

Since the development of next-generation sequencing (NGS) platform in 2005, NGS has revolutionized the fields of genetics and genomics.10 NGS technologies outperform conventional Sanger sequencing by dramatically decreasing cost and time of DNA sequencing. Target enrichment strategies including exome sequencing allow to effectively capture genomic regions of interest,11 and therefore, are useful to clinical sequencing in which DNA sequences of a large number of individuals are to be determined.

NGS technologies combined with target enrichment system are becoming an option for the identification of mutations in BRCA1 and BRCA2.12, 13, 14, 15, 16, 17, 18, 19, 20 According to the 2014 BRCA scheme report from the EMQN, 19% of laboratories use NGS technologies for BRCA testing.9 Some of these researches were conducted by commercial companies offering genetic tests for mutations including BRCA1 and BRCA2,14, 16, 17, 19, 20 especially after the U.S. Supreme Court in the June 2013 unanimously nullified Myriad Genetics’ patents on BRCA1 and BRCA2 (https://www.supremecourt.gov/opinions/12pdf/12-398_1b7d.pdf). The validity of NGS technologies were evaluated by comparing mutation status in patients with breast and ovarian cancers between NGS methods and conventional Sanger sequencing. It was demonstrated that NGS methods provided highly consistent results with conventional test with Sanger sequencing, and the sensitivity and specificity were close to perfect.12, 13, 14, 15, 16, 17, 18, 19, 20 Regardless of broad availability of genetic testing for mutations in BRCA1 and BRCA2 via NGS technologies, there are few applications to the general population.

In this study, we provide a ready-to-use protocol of identification for mutations in BRCA1 and BRCA2 via NGS combined with a target enrichment system. Our protocol realizes a rapid, high-quality, high-accuracy and low-cost sequencing by using very-low amounts of input DNA. A pool and capture method, where DNA libraries of 96 samples with distinct indices are pooled before hybridizing to the capture probes, largely reduce labor and cost. The proposed protocol showed that mutation statuses for 11 patients with breast cancer were concordant with those by the standard clinical testing. The application to 373 individuals in a general population identified novel stop gain substitution and frameshift deletion in BRCA2 in apparently unaffected individuals, suggesting the usefulness of large-scale screening for carriers of mutations in BRCA1/2. A future strategy for further cost reduction and extension to multigene panel are discussed.

Materials and methods

Study subjects

Participants of this study comprised 11 patients with breast cancer and 373 individuals in a general population who were resident of Chiba prefecture and received health check at Chiba University. The malignancies were pathologically confirmed. All the participants were of Japanese origin and recruited at Chiba University. All the participants underwent informed consent process for participation in research. The Ethics Committee of National Institute of Genetics and Chiba University approved the study protocols.

DNA samples

DNA samples of 11 patients were extracted from blood by using MabNA Pure Compact (Roche Diagnostics, Tokyo, Japan), and those of 373 individuals in the general population were from blood by using Gentra Puregene Blood Kit (Qiagen, Tokyo, Japan) according to the manufacturers’ protocols.

Genetic test for BRCA1 and BRCA2

Mutation statuses of BRCA1 and BRCA2 for 11 patients with breast cancer were determined by standard clinical testing (FALCO biosystems, Kyoto, Japan). It was reported that eight patients harbored nine mutations in these genes.

Design of target re-sequencing of BRCA1 and BRCA2

The target regions are designed to encompass all the transcripts of BRCA1 and BRCA2 based on UCSC Genes annotation (https://genome.ucsc.edu/index.html) and 5 kb regions flanking these genes (BRCA1, chr17: 41191312–41327420; and BRCA2, chr13: 32884596–32978809). A DNA probe set complementary to the target region was designed by NimbleDesign (https://design.nimblegen.com). Repetitive elements such as Alu elements constituting ~41.5% of the intronic regions of BRCA1 were masked.21

NGS library preparation combined with target enrichment system

The protocol introduced in this study is based on a pool and capture method by combining DNA library preparation with SureSelect QXT Library Prep Kit (Agilent Technologies, Santa Clara, CA, USA) with subsequent hybridization-based target enrichment with SeqCap EZ choice system (Roche Diagnostics).

We represent our protocol by organizing it into three sections. Unless otherwise stated, reagents used in sections (i), (ii), and (iii) are contained in SureSelect QXT Library Prep Kit (Agilent Technologies), SeqCap EZ choice system (Roche Diagnostics) and MiSeq Reagent Kit v3 (600 Cycles) (Illumina, San Diego, CA, USA), respectively.

i) DNA library preparation step (modified SureSelect QXT protocol)

1. Quantification of DNA concentration

  1. a

    Measure DNA concentration by Qubit dsDNA BR Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) on FilterMax F5 Multi-Mode Microplate Readers (Molecular Devices, Sunnyvale, CA, USA).

  2. b

    Adjust each DNA sample to a final concentration of 20 ng μl−1 by adding nuclease-free water.

2. Fragmentation and adapter ligation

  1. a

    Add 20 ng (1 μl) of DNA to well of a PCR plate on ice (up to 96 samples).

  2. b

    Add 1 μl of 0.5 × diluted SureSelect QXT Buffer.

  3. c

    Add 8 μl of SureSelect QXT Enzyme Mix ILM.

  4. d

    Seal the wells, mix thoroughly by vortex the plate for 20 s, and then spin down.

  5. e

    Incubate by a thermal cycler and run the following program:Step 1: 10 min at 45 °CStep 2: 1 min at 4 °CStep 3: hold at 4 °C

  6. f

    Add 16 μl of SureSelect Stop Solution to each well of the PCR plate on ice.

3. Purification

  1. a

    Incubate AMPure XP beads (Beckman Coulter, Fullerton, CA, USA) at room temperature (RT) for at least 30 min.

  2. b

    Add 16 μl of AMPure XP beads to each well, seal the wells with fresh caps, vortex for 5 s, and spin down.

  3. c

    Incubate at RT for 5 min.

  4. d

    Place on a magnetic stand for 5 min.

  5. e

    Discard all supernatant.

  6. f

    Wash 2 times as follows:Add 200 μl freshly prepared 80% ethanol.Incubate on the magnetic stand for 1 min.Discard all supernatant.

  7. g

    Dry up by the thermal cycler at 37 °C for 3 min with the lid open.

  8. h

    Add 11 μl of nuclease-free water to each well.

  9. i

    Seal the wells with fresh caps, mix by vortex, and spin down.

  10. j

    Incubate at RT for 2 min.

  11. k

    Place on a magnetic stand for 2 min.

  12. l

    Transfer 10 μl supernatant to a new PCR plate.

4. PCR amplification and addition of index tags

  1. a

    Prepare PCR reaction mix on ice. Volume for one reaction is as follows: Herculase II Reaction Buffer: 5 μlHerculase II Fusion DNA Polymerase: 0.5 μl100 mm dNTP Mix: 0.25 μlDMSO (100%): 1.25 μlPCR-grade water: 6 μl

  2. b

    Add 13 μl of PCR reaction mix to each sample.

  3. c

    Add 1 μl of P7 dual index primer.

  4. d

    Add 1 μl of P5 dual index primer.

  5. e

    Mix thoroughly by pipetting up and down.

  6. f

    Place the PCR plate on the thermal cycler and run the following PCR amplification program.Step 1: 2 min at 68 °CStep 2: 2 min at 98 °CStep 3: 30 s at 98 °CStep 4: 30 s at 57 °CStep 5: 1 min at 72 °CStep 6: 5 min at 72 °CStep 7: hold at 4 °CRepeat seven cycles of Step 3–5.

5. Purification

  1. a

    Incubate AMPure XP beads (Beckman Coulter) at RT for at least 30 min.

  2. b

    Add 25 μl of nuclease-free water.

  3. c

    Add 30 μl of AMPure XP beads to each well, seal the wells with fresh caps, vortex for 5 s, and spin down.

  4. d

    Incubate at RT for 5 min.

  5. e

    Place on a magnetic stand for 5 min.

  6. f

    Discard all the supernatant.

  7. g

    Wash 2 times as follows:Add 200 μl freshly prepared 80% ethanol.Incubate on the magnetic stand for 1 min.Discard all supernatant.

  8. h

    Dry up on the thermal cycler at 37 °C for 3 min with the lid open.

  9. i

    Add 11 μl of nuclease-free water to each well.

  10. j

    Seal the wells with new caps, mix by vortex, and spin down.

  11. k

    Incubate at RT for 2 min.

  12. l

    Place on the magnetic stand for 2 min.

  13. m

    Transfer 10 μl supernatant to a new PCR plate.

6. Pool samples

  1. a

    Measure DNA concentration by Qubit dsDNA BR Assay Kit (Thermo Fisher Scientific) on FilterMax F5 Multi-Mode Microplate Readers (Molecular Devices).

  2. b

    Pool equal amount of DNA from each sample to a final DNA amount of 1 μg into a new 1.5 ml tube.

ii) Hybridization step (modified SeqCap EZ protocol)

1. Hybridizing pooled DNA samples to probe set

  1. a

    Prepare hybridization enhancing (HE) oligo pool.

  2. b

    Add 2 μl (2 pmol) of HE oligo pool.

  3. c

    Add 5 μl of COT Human DNA (1 mg ml−1) (Sigma-Aldrich)

  4. d

    Evaporate by Centrifugal Concentrator CC-105 (TOMY Digital Biology).

  5. e

    Add a master mix of 7.5 μl of 2 × Hybridization Buffer and 3 μl of Hybridization Component A, and mix thoroughly by pipetting up and down.

  6. f

    Transfer 10.5 μl of pre-capture pooled DNA library to a new 0.2 ml tube.

  7. g

    Add 4.5 μl of SeqCap EZ probe pool, mix by vortex and spin down.

  8. h

    Place the PCR plate on the thermal cycler and run the following PCR amplification program.Step 1: 5 min at 95 °CStep 2: 20 h at 47 °C

2. Washing post-capture pooled DNA sample

  1. a

    Let stand Dynabeads M-270 Streptavidin (Thermo Fisher Scientific) for at least 45 min at RT.

  2. b

    Add 100 μl of Dynabeads to a new 1.5 ml tube.

  3. c

    Place on a magnetic stand, wait until the liquid becomes clear (5 min), and then remove and discard all the supernatant.

  4. d

    Wash 3 times as follows:Add 200 μl 1 × Beads Wash Buffer.Remove from the magnetic stand, and then mix by vortexing for 10 s.Place on the magnetic stand, wait until the liquid becomes clear, and then discard all the supernatant.

  5. e

    Transfer 15 μl of post-capture pooled DNA library to the beads containing tube, and mix by pipetting up and down.

  6. f

    Incubate on the thermal cycler at 47 °C for 45 min (with 3 s vortex in each 15 min).

  7. g

    Remove from the thermal cycler.

  8. h

    Add 100 μl of 1 × Wash Buffer I (incubated at 47 °C), and vortex for 10 s.

  9. i

    Place on a magnetic stand, wait until the liquid becomes clear, and then discard all the supernatant.

  10. j

    Wash two times as follows:Add 200 μl 1 × Stringent Wash Buffer (incubated at 47 °C), and mix by pipetting up and down for 10 times.Incubate on the thermal cycler at 47 °C for 5 min.Place on a magnetic stand, wait until the liquid becomes clear, and then remove and discard all supernatant.

  11. k

    Wash three times with distinct buffer and vortex time as follow:Add 200 μl 1 × buffer, and mix by vortex.Place on the magnetic stand, wait until the liquid becomes clear, and then remove and discard all supernatant.(first wash: Wash Buffer I and 2 min vortex; second wash: Wash Buffer II and 1 min vortex; and third wash: Wash Buffer III and 30 s vortex)

  12. l

    Add 20 μl PCR-grade water.

3. Amplifying post-capture pooled DNA library

  1. a

    Prepare master mix of 50 μl of 2 × KAPA HiFi HotStart ReadyMix (Kapa Biosystems), 2 μl of 100 μm TS-PCR oligo 1 & 2 (TS-PCR Oligo 1, 5′-AATGATACGGCGACCACCGAGA-3′; TS-PCR Oligo 2, 5′-CAAGCAGAAGACGGCATACGAG-3′), and 26 μl of PCR-grade water.

  2. b

    Add the master mix to 0.2 m tube with 20 μl of post-capture pooled DNA library.

  3. c

    Place the tube on the thermal cycler and running the following PCR amplification program.Step 1: 30 s at 98 °CStep 2: 10 s at 98 °C Step 3: 30 s at 60 °CStep 4: 30 s at 72 °CStep 5: 5 min at 72 °CStep 6: hold at 4 °CRepeat 12–18 cycles of Step 2–4.

  4. d

    Add 60 μl of AMPure XP beads, vortex for 5 s, and spin down.

  5. e

    Incubate at RT for 5 min.

  6. f

    Place on the magnetic stand for 5 min.

  7. g

    Discard all the supernatant.

  8. h

    Wash two times as follows:Add 200 μl freshly prepared 80% ethanol.Incubate on the magnetic stand for 1 min.Discard all supernatant.

  9. i

    Dry up on the thermal cycler at 37 °C for 3 min with the lid open.

  10. j

    Add 52 μl of nuclease-free water to each well.

  11. k

    Elute amplified DNA library for 2 min at RT.

  12. l

    Examine quality and quantity of amplified DNA library by DNA 1000 kit on Bioanalyzer 2100 and Qubit (BR).

iii) Sequencing on Illumina MiSeq

  1. a

    Dilute amplified DNA library to 4 nm.

  2. b

    Denature 5 μl of 4 nm DNA library by adding 5 μl of 0.2 N NaOH.

  3. c

    Mix by vortex, and centrifuge at 280 × g for 1 min.

  4. d

    Incubate for 5 min at RT.

  5. e

    Dilute DNA library to 20 pm by adding 990 μl of pre-chilled HT1.

  6. f

    Dilute DNA library to a final concentration (12 pm) by combining 360 μl of 20 pm DNA library and 240 μl of pre-chilled HT1.

  7. g

    Perform sequencing on the Illumina MiSeq platform with 350-bp and 250-bp paired-end module (Illumina).

  8. h

    Load pooled and diluted DNA library and custom sequencing primers (Read1, Index and Read2) provided in SureSelect QXT Library Prep Kit Box2 (Agilent Technologies) together with Illumina TruSeq Primers.

NGS data processing and variant calling

The reads containing the Illumina adapter sequences were trimmed by using Trimmomatic version 0.32.22 After the quality control step for excluding or trimming low quality sequences, the sequence reads were aligned to human reference genome (hg19) via BWA version 0.7.13.23 The information about the aligned reads was converted into compressed binary form (BAM format) and sorted based on genomic coordinates by using SAMtools version 1.3. (ref. 24) The aligned reads were processed for removal of PCR duplicates and erroneous reads by Picard tools version 1.136, and for local realignment and base quality recalibration by GATK version 3.2.2.25, 26 Coverages and average depths over target regions were calculated with the CallableLoci and DepthOfCoverage of the GATK, respectively.25, 26 Single nucleotide variants (SNVs) and insertions and deletions (indels) were detected with the HaplotypeCaller of the GATK.25, 26

Functional annotation of identified variants

SNVs and indels were classified into known and novel variants according to the presence in NCBI dbSNP build 138. Functional annotation was implemented via ANNOVAR version 1 February 2016.27 Estimates of mutation frequencies in East Asian populations were based on publicly available databases provided by following whole-genome and exome sequencing projects: the 1000 Genomes Project (1KG),28 the Exome Aggregation Consortium (ExAC),29 and the Human Genetic Variation Database (HGVD).30 The potential effects of the mutation were estimated by the online prediction and conservation tools: SIFT,31 PolyPhen2,32 and GERP++.33

Variant classification

Variants were classified for pathogenicity as follows: First, variants were classified as ‘benign’ if their frequencies in East Asian populations were greater than 1.0% in any of three publicly available databases: (1KG, ExAC and HGVD). Second, variants were classified as ‘pathogenic’ if they were loss-of-function variants (that is, nonsense, frameshift indels, canonical±1 or 2 splice sites or initiation codon). Third, we explored variants with previously established pathogenic or benign effects based on ClinVar.34 Finally, the remaining rare missense variants whose information about pathogenicity were not available on databases were classified into variants of uncertain significance (VUS).

Sanger sequencing

Sanger sequencing was performed using BigDye Terminator Cycle Sequencing V3.1 Ready Reaction kit (Life Technologies, Carlsbad, CA, USA) by forward or reverse primer on ABI 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). The oligonucleotides primers used are shown in Supplementary Table S1.

Results

We provide a ready-to-use protocol based on a pool and capture method by combining DNA library preparation with SureSelect QXT Library Prep Kit (Agilent Technologies) with subsequent hybridization-based target enrichment with SeqCap EZ choice system (Roche Diagnostics), in which DNA libraries are prepared with small amounts of input DNA (20 ng), up to 96 samples can be pooled before hybridization, and therefore only one hybridization reaction is required for 96 samples (Materials and Methods section). This modification permits saving a large amount of time and labor for performing hybridization reactions for all the samples and saving costs for reagents for hybridization. Our protocol comprises three sections: DNA library preparation, hybridization, and sequencing steps. They take 7.5, 25.5 and 55.5 h, respectively (Figure 1). Therefore, we can obtain targeted sequences within 5 days.

Figure 1
figure 1

Workflow and processing time of protocol of sequencing for BRCA1 and BRCA2. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

We selected a total of 230 kb regions encompassing all the transcripts of BRCA1 and BRCA2 based on UCSC Genes annotation and 5 kb regions flanking these genes (Figure 2). It has been reported that ~41.5% of the intronic regions of BRCA1 consist of Alu elements (Figure 2).21 We did not design probes on such repetitive elements to avoid false variant calls.

Figure 2
figure 2

Design of target enrichment system for (a) BRCA1 and (b) BRCA2. In each panel, all the transcripts based on UCSC Genes annotation, repeated sequences based on RepeatMasker from UCSC genome browser, and probe set are depicted. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

In order to evaluate sequence quality of our protocol, we evaluated sequence data from 384 individuals regarding coverage and average depth over coding exons and their flanking 20 bp regions according to the fact that most pathogenic mutations in BRCA1 and BRCA2 are predicted to produce truncated proteins (that is, nonsense and splice site mutations, and frameshift indels). Greater than 99.99% of the coding exons and their flanking 20 bp regions were covered at least 20 reads, and the average depth was 179.5. This result indicates that the method provides almost complete coverage and confidential variant calling for all the 384 individuals with limited amounts of input DNA.

We verified the accuracy of our variant detection and classification by comparing results of 11 patients with breast cancer whose mutation statuses had been determined by standard clinical testing. According to results of genetic tests provided by a company, eight of these 11 patients carried nine notable mutations: Two nonsense mutations in BRCA1 (c.T188A [p.L63X], c.C3607T [p.R1203X]), one frameshift insertion in BRCA1 (c.66dupA [p.E23fs]), and one frameshift deletion in BRCA2 (c.6597_6598del [p.T2199fs]) were reported to be pathogenic. One splice site variant in BRCA1 (c.4485-2A>G) was considered as likely pathogenic. A total of four missense mutations in BRCA1 (c.C626T [p.P209L], c.A2726T [p.N909I]) and BRCA2 (c.A3395G [p.K1132], c.T3420A [p.S1140R]) were categorized into VUS (Table 1). The proposed sequence method successfully identified all of these mutations. The results of our variant classification were completely consistent with those of the standard clinical testing.

Table 1 Mutations identified in 8 patients with breast cancer
Full size table

We determined sequences of BRCA1 and BRCA2 for 373 individuals from the general Japanese population. We detected 76 variants on coding exons (Supplementary Table S2). Twenty seven synonymous variants and 16 variants were predicted to be benign according to their functional significance and frequencies in East Asian populations showing greater than or equal to 1%. We identified one nonsense mutation (c.A6922T [p.K2308X]) and one frameshift deletion (c.3571delA [p.K1191fs]) in BRCA2, which were predicted to be pathogenic (Table 2). All these pathogenic variants were confirmed by Sanger sequencing (Supplementary Figure S1). The estimated proportion of harboring germline deleterious mutations in BRCA1 or BRCA2 was 0.54% (95% confidence interval, 0.065–1.9%). Additionally, one non-frameshift deletion in BRCA2 categorized into VUS was detected (c.9106_9108del [p.3036_3036del]; Table 2). The discovery of deleterious variants in the general population highlights the importance of screening of BRCA1 and BRCA2 mutation carriers for early detection and prevention.

Table 2 Mutations identified in 3 individuals from a general Japanese population
Full size table

Although current genetic testing for mutations in BRCA1 and BRCA2 mainly focus on the coding sequences and exon–intron boundaries, researchers has reported that an intronic mutation (c.6937+594T>G) in BRCA2 led to aberrant transcription of a cryptic exon, which resulted in degradation of mutant transcript by nonsense mediated decay.35 We did not observe this intronic mutation in this study. There is a possibility that exploration for intronic mutations can be useful for patients and families without coding mutations.35 Our target enrichment system was designed to capture intronic regions of BRCA1 and BRCA2 except for repetitive regions. Therefore, we determined intronic sequences and provided information about allele frequencies in the general population (Supplementary Table S3), which may be helpful for filtering benign variants according to their frequencies.

Discussions

Since mutations in BRCA1 and BRCA2 were discovered in the middle 1990s in families with breast and ovarian cancer, genetic testing for mutations in these genes has become more versatile tool in clinical genetics.1 Genetic testing was first used to women with cancer diagnosis at a young age and with a substantial family history, bilateral breast cancer, or both breast and ovarian cancer. Cumulative evidence of decreases in cancer incidence, morbidity and mortality by surgical intervention, it has now expanded to cancer-free women.3, 4, 5, 6

The innovation in sequencing technology has enabled an unprecedented amount of sequence data at low cost.10 Altogether with target enrichment system and multiplexed indices, genetic regions of interest can be simultaneously sequenced for a large number of samples in a cost-effective fashion.11, 36 Target enrichment systems used for sequencing of cancer predisposing genes including BRCA1 and BRCA2 can be roughly grouped into two categories: PCR-based13, 14, 16, 37, 38, 39, 40, 41 and hybridization-based15, 17, 18, 19, 20, 42, 43, 44 approaches. There are several pros and cons to be considered on these two approaches. Pooling of typical PCR amplicons designed to amplify all coding exons and exon–intron boundaries seems to be too laborious and consumes large amounts of DNA when numbers of exons and/or samples are large. Multiplex PCR usually generates nonspecific amplification due to interaction between primers. Most of the current PCR-based approaches rely on technologies using microdroplets, in which each droplet exerts an independent PCR reaction followed by sequencing the mixture of DNA amplicons from many emulsions.45 The hybrid capture is well established, in which a fragmented library is hybridized to specific probes designed to target regions of interest. The advantage of the hybridization method is capability of capturing large target regions in a single experiment with limited amounts of input DNA. On the other hand, the specificity of PCR-based method outperforms hybridization-based method, in which probes to repetitive regions cannot be designed to reduce nonspecific hybridization. A potential drawback of PCR-based approach is low-throughput and allele dropout due to rare sequence variants under primer binding sites.14, 16

In this study, we provide a ready-to-use protocol based on a pool and capture method by combining DNA library preparation with SureSelect QXT Library Prep Kit with subsequent hybridization-based target enrichment with SeqCap EZ choice system. The cost of performing target enrichment by hybridization is reduced by pooling samples before hybridization, in which only one hybridization reaction is required for 96 samples. Similar approach is proposed, in which 96 distinct indices are used to distinguish samples at the sequencing step but at most 12 samples are hybridized to probes of SeqCap EZ choice system in a single enrichment reaction.46 We demonstrated that 96 samples could be hybridized in a single enrichment reaction, which will largely reduce the cost. It is notable that the sequencing cost for one individual is less than 8000 JPY (80 USD) even if we calculate the cost for all the requirements based on catalog price. The proposed protocol is easy-to-follow for laboratory workers with basic technology in molecular biology. The workflow from DNA to sequences completes in five days. Though majority of the 5 days are waiting time for hybridization and sequencing steps. The entire steps including bioinformatics analyses can be handled within a week, which allows medical staff to take enough time to perform variant classification.

In this study, we applied the method for the identification of mutations in BRCA1 and BRCA2. We confirmed that all the mutations in patients with breast cancer were identified and classified consistently with the results of standard clinical testing. Then, we extended our analysis to subjects participated in a general health examination. Among the 373 general individuals, we identified two pathogenic mutations produce truncated proteins. The estimated proportion of harboring germline deleterious mutations in BRCA1 or BRCA2 was 0.54% (95% confidence interval, 0.065–1.9%). This mutation prevalence is comparable with that in non-Jewish populations (~0.25%) but lower than that in Ashkenazi-Jewish (~2.5%).47, 48 Although further evaluations are required, cost-effectiveness of population-based screening for BRCA1 and BRCA2 will be improved by the achievement of reducing the cost. Furthermore, women who want to be tested for mutations in these genes might defray their own costs if all expenses related to the test decrease to the same level as achieved in this study.

The logical extension of our protocol is to develop multigene screening panels. We can easily construct multigene screening panels at the same cost by including exons and their flanking regions instead of introns of BRCA1 and BRCA2 that are not successfully captured in the current probe design. Pooling larger number of samples and subsequent sequencing via higher throughput sequencer (for example, HiSeq) will enable to further reduce the cost for multigene screening panels. This protocol can be applied to detect somatic mutations, which may be clinically important because germline and somatic mutations in DNA-repair genes including BRCA1/2-affected therapeutic response rate to an anticancer drug in patients with prostate cancer.49 Another extension is to establish a sequencing method to fill the gaps in the intronic regions containing Alu repeats. This is important for determination of breakpoints of large insertions and deletions flanking by Alu sequences that mediate the mutation.12 Our protocol can be flexibly applied to a wide variety of target re-sequencing studies such as fine mapping of disease associated loci50 and determination of human leukocyte antigen genes.51