"Sequencing-grade" screening for BRCA1 variants by oligo-arrays
© Monaco et al; licensee BioMed Central Ltd. 2008
Received: 15 August 2008
Accepted: 30 October 2008
Published: 30 October 2008
The need for fast, efficient, and less costly means to screen genetic variants associated with disease predisposition led us to develop an oligo-nucleotide array-based process for gene-specific single nucleotide polymorphism (SNP) genotyping. This cost-effective, high-throughput strategy has high sensitivity and the same degree of accuracy as direct sequencing, the current gold standard for genetic screening. We used the BRCA1 breast and ovarian cancer predisposing gene model for the validation of the accuracy and efficiency of our strategy. This process could detect point mutations, insertions or deletions of any length, of known and unknown variants even in heterozygous conditions without affecting sensitivity and specificity. The system could be applied to other disorders and can also be custom-designed to include a number of genes related to specific clinical conditions. This system is particularly useful for the screening of long genomic regions with relatively infrequent but clinically relevant variants, while drastically cutting time and costs in comparison to high-throughput sequencing.
High throughput $1,000 whole genome sequencing may be rapidly approaching[1, 2], meanwhile, a clinical need exists for the screening of genes whose polymorphisms determine disease predisposition, natural history or therapeutic outcome. Screening of the BRCA1 (OMIM 113705) cancer predisposition genes is an example of such a situation and it was well exemplified by [3, 4] by Gerhardus et al , who systematically reviewed 3816 publications to estimate the accuracy of diagnostic methods used for the detection of BRCA1 and BRCA2 mutations. They concluded that many of the alternative screening methods were as time- and cost-intensive as direct sequencing, but did not provide the same definitive information. In addition, many of these methods could not be recommended for routine screening because of low sensitivity. Denaturing high-performance liquid chromatography was shown to outperform other methods but still required to be complemented by sequencing. Significantly, none of the techniques evaluated in the study, including direct sequencing, could detect large rearrangements, such as whole exon germline deletions/insertions.
Results and discussion
Estimated cost and time requirements for typing of the BRCA1 gene by direct sequencing vs SNP array
Total Cost/sample BRCA1 gene (35 fragment)
approx 2 working days
approx 20 working days
less than 3 working days
less than 3 working days
To evaluate the sensitivity and specificity of the process, we compared results obtained with the oligoarray against those from direct sequencing. As part of an ongoing clinical protocol, samples for BRCA1 and BRCA2 mutational analysis were obtained from 85 consecutive patients with familial breast and/or ovarian cancer. Patients were seen and signed informed consent at the Genetic Counselling Program, Clinical Oncology Laboratory, at the Bari National Cancer Institute (DNV Certificate N. CERT-17885-2006-AQ-BRI-SINCERT). Only patients classified as having a higher than 10% probability of carrying a BRCA1 or BRCA2 mutation were enrolled. This risk was calculated using the New Myriad II program, which references an individual's TNM classification U.I.C.C., cyto-histological differentiation grade, estrogen receptor (ER) and progesterone receptor (PgR) status, tumor content and whether there is a history of breast or ovarian cancer among relatives.
Fragment 4 of the BRCA1 locus contains several SNPs associated with the predisposition for developing breast and ovarian cancer and was used for SNP analysis and validation by direct sequencing (Figure 2). To demonstrate the principle, data were portrayed for individual fragments (sub-arrays) after fragment-specific normalization to graphically display the presence of SNPs along the sequence (as previously described; Figure 2). Consistent calls identifying SNPs present in the reference sample (that was not completely identical to the wild-type consensus sequence) in all cases were excluded from the analysis because representative of variations in the reference MCF-10A cell line and not related to the test sample. This fragment-specific normalization corrects sequence-specific and amplicon-specific variation in intensity that may cause imbalanced hybridization as tested using sequence identical samples differentially labeled and hybridized on the same chip for calibration purposes (see example in Figure 2, top panel). This normalization does not affect the intra-fragment reference/test ratio measurements.
A custom made software SNPpositioner uses an algorithm that queries the Graphical User Interface to select pre-determined chromosomal regions relevant to the analysis (individual fragments in this case). Probe logRatio were first averaged from duplicated spots followed by the "Local Amplicon-oriented Normalization Algorithm" (LANA). This LANA approach is used to sort individual probes implementing the two nearest flanking probes summarized below:
T(log Ratio i ) = 2*log Ratio i - log Ratioi-1- log Ratioi+1
Data analysis, therefore, is performed blindly and automatically to identify variant sequences when the transformed Cy5/Cy3 logRatio [T(logRatio)] of a probe is above and/or below a fragment-defined baseline cutoff value which is two standard deviations in the current settings. This algorithm objectively identifies sequence variations without any subjective manipulation the oligo-array data. The analysis was carried out blind (only the reference was completely sequenced for the BRCA1 locus) and it was automated using our custom software that made calls without input from previous sequencing information. Thus, the study was used as a training set for the program.
To ensure the accuracy of this technology and analysis software, the output SNP information was compared with sequence-based analysis of 2 kilobases region in fragment 4 (Figure 3A). This comparison identified complete concordance between SNPs identified by SNPpositioner and those made by sequencing analysis for 83 of the 85 patient samples (highlighted in yellow in Figure 3A; two patient samples could not be sequenced due to insufficient DNA and, therefore, the accuracy of the array could not be tested in those). In these 85 patients, the oligo-array detected 15 non-synonymous, 4 synonymous and 10 intronic SNPs. No novel SNPs were identified in this previously well-characterized Italian population [4, 6, 13]. In about 50% of patients tested, three SNPs (P871L, K1183R and E1038G) were consistently present, indicating possible haplotype linkage. When cross-referenced with clinical-pathological information, these three linked SNPs identified a cluster of individuals possessing a higher percentage of cyto-histologically differentiated cancers as compared with the other patients (71% [27/38] vs 50% [19/38] of G1-2 tumours; p = 0.05). These patients also had a lower probability of carrying a deleterious BRCA1 or BRCA2 mutation (74% [31/42] vs 56% [24/43] of cases with Myriad probability ≤ 10%, p = 0.06) .
Although the oligo-array's accuracy was only confirmed with sequencing by fragment 4 of the BRCA1 locus, it could be expected that the same accuracy would be observed with other fragments. Thus, the whole BRCA1 gene can be analyzed with one oligo-array reaction and have the same accuracy as at least 70 sequencing reactions (about 35 kb). In addition, the automated data interpretation eliminated regions of balanced hybridization limiting the analysis to only those few regions flagged by the software to contain SNPs, therefore, greatly simplifying the analysis. A comparative analysis of the time and cost of the two techniques is shown in Table 1. Our estimates of the cost of sequencing for the BRCA1 were similar to others' reports .
In summary, the process presented here is an accurate and efficient screening strategy for gene-specific detection of clinically or scientifically relevant genomic variants. This validation should be regarded as a further improvement in the efficiency of genetic testing as discussed by Gerhardus et al . Contrary to previously sequencing-on-chip methods [14–18], this method can detect known gene variants  with high sensitivity while using a much smaller number of oligos. Indeed, other systems comparable to the present in potential accuracy such as "on chip sequencing" cover a complete gene sequence tiling oligos with a 1 nucleotide overlap and including probes for each possible nucleotide permutation for each base position. This study clearly shows that for practical purposes, such as clinical-grade genetic testing, this extensive approach is not necessary and wasteful; in fact, although in theory it eliminates the requirement for sequencing, in practice it requires a large number of oligos to cover areas that are in most cases non-polymorphic or test genes whose polymorphisms are in most cases known (as the BRCA1 gene). Our process can theoretically flag the occurrence of unknown variants based on the sequential signal loss pattern in tiled consensus oligo probes, although not tested in this study in which well-characterized patients were screened; in this case sequencing in the search for new variant sequences could be focused to extremely limited areas in rare patients (only those patients carrying novel SNPs). We estimate that this process could reduce the need for direct sequencing to less than the 1% of present norms. In addition, because of the small number of oligos needed, as compared with sequencing-on-chip technologies, this strategy dramatically reduces the production costs. It may also allow the inclusion of several genes relevant to a specific disease process to be analyzed simultaneously at "sequence-grade" levels using high-density platforms.
Tyler Pierson, Brunella Pilato, Rosanna Lacalamita, Rosamaria Pinto, Andrea Worschech, Zoltan Pos.
- Mardis ER: Anticipating the 1,000 dollar genome. Genome Biol. 2006, 7: 112-10.1186/gb-2006-7-7-112.PubMed CentralView ArticlePubMedGoogle Scholar
- Mardis ER: ChIP-seq: welcome to the new frontier. Nat Methods. 2007, 4: 613-614. 10.1038/nmeth0807-613.View ArticlePubMedGoogle Scholar
- Antoniou AC, Pharoah PD, McMullan G, Day NE, Stratton MR, Peto J, Ponder BJ, Easton DF: A comprehensive model for familial breast cancer incorporating BRCA1, BRCA2 and other genes. Br J Cancer. 2002, 86: 76-83. 10.1038/sj.bjc.6600008.PubMed CentralView ArticlePubMedGoogle Scholar
- Tommasi S, Crapolicchio A, Lacalamita R, Bruno M, Monaco A, Petroni S, Schittulli F, Longo S, Digennaro M, Calistri D, Mangia A, Paradiso A: BRCA1 mutations and polymorphisms in a hospital-based consecutive series of breast cancer patients from Apulia, Italy. Mutat Res. 2005, 578: 395-405.View ArticlePubMedGoogle Scholar
- Gerhardus A, Schleberger H, Schlegelberger B, Gadzicki D: Diagnostic accuracy of methods for the detection of BRCA1 and BRCA2 mutations: a systematic review. Eur J Hum Genet. 2007, 15: 619-627. 10.1038/sj.ejhg.5201806.View ArticlePubMedGoogle Scholar
- Bruno M, Tommasi S, Stea B, Quaranta M, Schittulli F, Mastropasqua A, Distante A, Di Paola L, Paradiso A: Awareness of breast cancer genetics and interest in predictive genetic testing: a survey of a southern Italian population. Ann Oncol. 2004, 15 (Suppl 1): I48-I54. 10.1093/annonc/mdh658.View ArticlePubMedGoogle Scholar
- Narod SA, Foulkes WD: BRCA1 and BRCA2: 1994 and beyond. Nat Rev Cancer. 2004, 4: 665-676. 10.1038/nrc1431.View ArticlePubMedGoogle Scholar
- Sevilla C, Moatti JP, Julian-Reynier C, Eisinger F, Stoppa-Lyonnet D, Bressac-de Paillerets B, Sobol H: Testing for BRCA1 mutations: a cost-effectiveness analysis. Eur J Hum Genet. 2002, 10: 599-606. 10.1038/sj.ejhg.5200854.View ArticlePubMedGoogle Scholar
- Wang E, Adams S, Zhao Y, Panelli M, Simon R, Klein H, Marincola FM: A strategy for detection of known and unknown SNP using a minimum number of oligonucleotides. J Transl Med. 2003, 1: 4-10.1186/1479-5876-1-4.PubMed CentralView ArticlePubMedGoogle Scholar
- Marmur J, Doty P: Determination of the base composition of deoxyribonucleic acid from its thermal denaturation temperature. J Mol Biol. 1962, 5: 109-118.View ArticlePubMedGoogle Scholar
- Breslauer KJ, Frank R, Blocker H, Marky LA: Predicting DNA duplex stability from the base sequence. Proc Natl Acad Sci USA. 1986, 83: 3746-3750. 10.1073/pnas.83.11.3746.PubMed CentralView ArticlePubMedGoogle Scholar
- Sugimoto N, Nakano S, Yoneyama M, Honda K: Improved thermodynamic parameters and helix initiation factor to predict stability of DNA duplexes. Nucleic Acids Res. 1996, 24: 4501-4505. 10.1093/nar/24.22.4501.PubMed CentralView ArticlePubMedGoogle Scholar
- Tommasi S, Fedele V, Lacalamita R, Bruno M, Schittulli F, Ginzinger D, Scott G, Eppenberger-Castori S, Calistri D, Casadei S, Seymour I, Longo S, Giannelli G, Pilato B, Simone G, Benz CC, Paradiso A: 655Val and 1170Pro ERBB2 SNPs in familial breast cancer risk and BRCA1 alterations. Cell Oncol. 2007, 29: 241-248.PubMedGoogle Scholar
- Patil N, Berno AJ, Hinds DA, Barrett WA, Doshi JM, Hacker CR, Kautzer CR, Lee DH, Marjoribanks C, McDonough DP, Nguyen BT, Norris MC, Sheehan JB, Shen N, Stern D, Stokowski RP, Thomas DJ, Trulson MO, Vyas KR, Frazer KA, Fodor SP, Cox DR: Blocks of limited haplotype diversity revealed by high-resolution scanning of human chromosome 21. Science. 2001, 294: 1719-1723. 10.1126/science.1065573.View ArticlePubMedGoogle Scholar
- Chee M, Yang R, Hubbell E, Berno A, Huang XC, Stern D, Winkler J, Lockhart DJ, Morris MS, Fodor SP: Accessing genetic information with high-density DNA arrays. Science. 1996, 274: 610-614. 10.1126/science.274.5287.610.View ArticlePubMedGoogle Scholar
- Hacia JG, Sun B, Hunt N, Edgemon K, Mosbrook D, Robbins C, Fodor SP, Tagle DA, Collins FS: Strategies for mutational analysis of the large multiexon ATM gene using high-density oligonucleotide arrays. Genome Res. 1998, 8: 1245-1258.PubMedGoogle Scholar
- Hacia JG: Resequencing and mutational analysis using oligonucleotide microarrays. Nature Genetics. 1999, 21: 42-47. 10.1038/4469.View ArticlePubMedGoogle Scholar
- Murabito JM, Rosenberg CL, Finger D, Kreger BE, Levy D, Splansky GL, Antman K, Hwang SJ: A genome-wide association study of breast and prostate cancer in the NHLBI's Framingham Heart Study. BMC Med Genet. 2007, 8 (Suppl 1): S6-10.1186/1471-2350-8-S1-S6.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.