24 Apr 2014: (2014) Correction: Next Generation MUT-MAP, a High-Sensitivity High-Throughput Microfluidics Chip-Based Mutation Analysis Panel. PLoS ONE 9(4): e96019. doi: 10.1371/journal.pone.0096019 | View correction
Molecular profiling of tumor tissue to detect alterations, such as oncogenic mutations, plays a vital role in determining treatment options in oncology. Hence, there is an increasing need for a robust and high-throughput technology to detect oncogenic hotspot mutations. Although commercial assays are available to detect genetic alterations in single genes, only a limited amount of tissue is often available from patients, requiring multiplexing to allow for simultaneous detection of mutations in many genes using low DNA input. Even though next-generation sequencing (NGS) platforms provide powerful tools for this purpose, they face challenges such as high cost, large DNA input requirement, complex data analysis, and long turnaround times, limiting their use in clinical settings. We report the development of the next generation mutation multi-analyte panel (MUT-MAP), a high-throughput microfluidic, panel for detecting 120 somatic mutations across eleven genes of therapeutic interest (AKT1, BRAF, EGFR, FGFR3, FLT3, HRAS, KIT, KRAS, MET, NRAS, and PIK3CA) using allele-specific PCR (AS-PCR) and Taqman technology. This mutation panel requires as little as 2 ng of high quality DNA from fresh frozen or 100 ng of DNA from formalin-fixed paraffin-embedded (FFPE) tissues. Mutation calls, including an automated data analysis process, have been implemented to run 88 samples per day. Validation of this platform using plasmids showed robust signal and low cross-reactivity in all of the newly added assays and mutation calls in cell line samples were found to be consistent with the Catalogue of Somatic Mutations in Cancer (COSMIC) database allowing for direct comparison of our platform to Sanger sequencing. High correlation with NGS when compared to the SuraSeq500 panel run on the Ion Torrent platform in a FFPE dilution experiment showed assay sensitivity down to 0.45%. This multiplexed mutation panel is a valuable tool for high-throughput biomarker discovery in personalized medicine and cancer drug development.
Citation: Schleifman EB, Tam R, Patel R, Tsan A, Sumiyoshi T, et al. (2014) Next Generation MUT-MAP, a High-Sensitivity High-Throughput Microfluidics Chip-Based Mutation Analysis Panel. PLoS ONE 9(3): e90761. doi:10.1371/journal.pone.0090761
Editor: Jörg D. Hoheisel, Deutsches Krebsforschungszentrum, Germany
Received: November 8, 2013; Accepted: February 3, 2014; Published: March 21, 2014
Copyright: © 2014 Schleifman et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was funded by Genentech, Inc. The funders were responsible for the study design, data collection and analysis, decision to publish, and preparation of the manuscript.
Competing interests: The authors have read the journal's policy and have the following conflicts: All studies were funded by Genentech, Inc. Erica Schleifman, Rachel Tam, Rajesh Patel, Teiko Sumiyoshi, Ling Fu, Rupal Desai, and Rajiv Raja are or were employed by Genentech, and Alison Tsan, Nancy Schoenbrunner, Thomas W. Myers, Keith Bauer, and Edward Smith are or were employed by Roche Molecular Systems Inc. There are no patents, products in development or marketed products to declare. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Biological markers, or biomarkers, have been defined as “any substance, structure or process that can be measured in bio-specimen and which may be associated with health-related outcomes” . Currently biomarkers are being used for prognostic, diagnostic, and predictive purposes in the field of oncology and as such play a vital role in personalized medicine. Biomarkers can be used to determine subsets of a population that may or may not respond to drug treatment/therapy and can even be used to prescreen patients in clinical trials. The reliable detection and validation of these markers is therefore essential.
In the last ten years, developments in genome-wide analytic methods have made the profiling of gene expression and genetic alternations of the cancer genome possible. By determining the molecular profile of a tumor (both mutational status and gene expression), a patient's disease can be characterized. This information can then be used to determine which course of treatment a patient should follow. A recent example of such targeted therapy is the development of ZELBORAF for treatment of patients whose unresectable or metastatic melanoma harbors a BRAF V600E mutation . A companion diagnostic assay was developed with this drug to screen patients, allowing only those patients whose tumors were biomarker positive to receive the treatment. Somatic mutations, therefore, can serve as tumor specific biomarkers, allowing for the use of targeted therapies.
One of the biggest challenges in using clinical samples for biomarker detection is the fact that most tumor biopsies are formalin-fixed and paraffin-embedded (FFPE) for long term storage of the tissue . This treatment leads to lower yield and quality of isolated genomic DNA (gDNA) from the samples due to cross-linking and fragmentation.
Characterization of the cancer genome by next generation sequencing (NGS) methods have emerged, ignited by the increased understanding of somatic alternations in cancer and their value in the development of personalized therapeutics. However, NGS lacks the analytical sensitivity and quantitative performance required for mutation detection in FFPE tissues. Furthermore, currently NGS requires larger DNA quantities for analysis, has complex and time consuming data analysis pipelines, and involves high costs, all of which makes NGS impractical for routine clinical use.
We previously developed a mutation multi-analyte panel (MUT-MAP) that allowed for the detection of 71 mutations across six oncogenes. This panel utilized the Fluidigm microfluidics technology which allowed for simultaneous detection of these mutations in a single sample. We report here the development and validation of the next generation MUT-MAP, a high-throughput platform that can now detect 120 hotspot mutations in eleven genes (AKT1, BRAF, EGFR, FGFR3, FLT3, HRAS, KIT, KRAS, MET, NRAS, and PIK3CA) based on allele specific PCR (AS-PCR) and Taqman technologies. Analysis of 88 samples can be completed in one day with as little as 2 ng of high quality gDNA or 100 ng of gDNA derived from FFPE tissues. By multiplexing our assays, less precious sample is required, resulting in a robust and easy to interpret data output.
The mutations detected in this panel are found in various types of cancers and the genes encode proteins of therapeutic interest. For example, bladder cancer has a 44% frequency of mutations in FGFR3,13% in RAS oncogenes (HRAS, KRAS, or NRAS), and 13–27% in PIK3CA. These mutations are currently being validated as potential diagnostic biomarkers for patient stratification in clinical trials . Mutations in FLT3 lead to constitutively active FLT3 which can then act in a ligand-independent manner in leukemia . KIT mutations have been implicated in several cancers including melanoma  and gastrointestinal stromal tumors . MET mutations are prevalent in hereditary and sporadic papillary renal cell carcinoma , head and neck carcinoma , and non-small cell and small cell lung cancer .
The updated MUT-MAP microfluidics system continues to provide a cost-effective, high-sensitivity, and high-throughput platform for exploratory analysis of predictive and prognostic biomarkers in clinical trial samples. It offers a means of detecting a wide range of mutations in a panel of eleven therapeutically relevant genes. The MUT-MAP system reported here can be used to analyze somatic mutations with very small amounts of gDNA from poor quality, archived FFPE tissues and could be used for exploratory biomarker analysis supporting the development of tools for predictive and prognostic assessment of various cancers.
Materials and Methods
The updated MUT-MAP panel was run on the BioMark platform (Fluidigm Corp.) using a 96.96 dynamic array as described previously  with a few alterations. Preamplified DNA combined with qPCR reagents and 10× assays mixed with the Fluidigm 20× sample loading reagent (Fluidigm Corp.) were loaded onto the chip as per the manufacturer's protocol. All newly added assays were allele-specific PCR (AS-PCR) assays which utilized an engineered Thermus specie Z05 DNA polymerase (AS1) and primers to allow for allelic discrimination between the wild-type and mutant sequence. ,  An exon specific probe was used in all assays.
DNA preamplification procedures were performed as described previously . Briefly, DNA was preamplified in a 10 µl reaction for 20 cycles in the presence of a preamplification primer cocktail mix (Table S1 shows sequences of newly added primers) and 1× ABI Preamp Master Mix (Applied Biosystems; Foster City, CA). All samples were exonuclease treated after PCR amplification to remove the remaining primers before being loaded onto the chip. Exonuclease I (16 U) (New England Biolabs; Ipswitch, MA) in exonuclease reaction buffer and nuclease-free water were added to each 10 µl PCR amplification and incubated at 37°C for 30 min followed by a 15 min incubation at 80°C for enzyme inactivation. Samples were then diluted four-fold in nuclease-free water and stored at 4°C or −20°C until needed.
A positive control was prepared in bulk by amplification of a cocktail of relevant mutant plasmids for all eleven genes in the presence of a wild-type human genomic DNA background; this positive control was run in triplicate on every chip for quality control purposes.
Preparation of Reagents
All assays from the previous MUT-MAP were prepared as described previously . Final primer and probe concentrations of 200 and 100 nM were used respectively for the newly designed custom AS-PCR assays which were added to the panel. These assays are currently under development at Roche Molecular Systems, Inc. (Pleasanton, CA).
A commercially available COBAS PIK3CA Mutation Test (Roche Molecular Systems) was modified to achieve compatibility with the two-color BioMark readout (FAM and VIC) for mutation detections in the PIK3CA gene.
All assays were prepared by diluting assays with the 20× sample loading buffer (Fluidigm Corp.). Diluted samples were mixed with AS1 qPCR master mix and run in duplicate by loading 5 µL into each well of a primed 96.96 Fluidigm Chip. The 96.96 dynamic array was loaded and then analyzed with the BioMark reader as previously described .
Data was analyzed and cycle threshold (CT) values were determined using the BioMark real-time PCR analysis software (Fluidigm Corp.) and automated mutation calls were determined using an algorithm based on the difference in CT (ΔCT) values between wild-type and mutant assays for all AS-PCR assays.
Eleven-Gene Mutation Panel
This MUT-MAP panel can screen 120 hotspot mutations across the AKT1, BRAF, EGFR, FGFR3, FLT3, HRAS, KIT, KRAS, MET, NRAS, and PIK3CA genes. The mutation coverage of additional content on this panel is presented in Table 1.
Table 1. Mutation Coverage Breakdown by Gene.doi:10.1371/journal.pone.0090761.t001
Assay Specificity and Sensitivity
Individual plasmids, each containing a single mutation correlating to each newly added assay on the 11-gene panel were used as samples to determine assay specificity and determine potential cross-reactivity between different hotspots.
Five linearized mutant plasmids were mixed to a final concentration of 4 ng/µL. The resulting mixes were diluted in either nuclease-free water or wild-type genomic DNA (Taqman Control Human Genomic DNA, Life Technologies, Cat# 4312660) where the genomic DNA concentration was kept constant at 10 ng. All of the samples were analyzed by the 11-gene panel along with a standard curve of wild-type human gDNA alone. Percentage of each mutation detected was calculated and the lower limit of detection (LLOD) of the assays in a genomic DNA background was determined for each assay evaluated. The samples diluted in nuclease-free water allowed for the assessment of assay linearity.
Mutation calls were validated using cell lines as well as FFPE tissues. Cell lines with known mutations reported in the literature were used to confirm the sensitivity and specificity of the assays. Further, a total of nine FFPE samples with known mutation status were mixed together with varying DNA inputs into seven Latin square mixes. The final DNA concentration of each mix was 40 ng/µl. These seven mixes were analyzed on MUT-MAP as well as by the SuraSeq500 panel on the Ion Torrent platform  in order to compare mutation calls and sensitivity levels of both platforms. The resulting data has been uploaded to the European Nucleotide Archive, http://www.ebi.ac.uk/ena/data/view/PRJEB5209.
To increase the coverage of our MUT-MAP platform, AS-PCR assays for HRAS, FGFR3, FLT3, KIT, MET, and PIK3CA were added (Table 1). The updated panel can now detect 120 somatic mutations across eleven genes of therapeutic interest for a single sample. By multiplexing assays and using two detection channels (FAM and VIC), we were able to consolidate all the assays onto a single Fluidigm microfluidics chip allowing for the simultaneous detection of 120 mutations in 44 samples.
Mutant Control Formulation
A single control sample was formulated to be used as a positive control for every assay on MUT-MAP using the process described in Figure 1A. The positive control was generated by mixing mutant plasmids in the presence of a wild-type human genomic DNA background. The positive control was further preamplified and diluted to a concentration that resulted in CT ranges from 9–16 across all wild-type and mutant assays (Figure 1B). This mutant control is included in every chip for quality control purposes.
(A) Schematic diagram for the process of generating the positive control for MUT-MAP. (B) The positive control is a mixture of mutant plasmids and wild-type human genomic DNA. The positive control was created such that the resulting CTs range from 9–16 across all wild-type and mutant assays. Pk_H1047X covers multiple hotspot mutations resulting in a lower overall CT as it is detecting more than one plasmid in the positive control.doi:10.1371/journal.pone.0090761.g001
A series of experiments were performed to validate the new assays added to the panel to ensure specificity and reproducibility. As described previously , a complete cross-reactivity analysis was conducted by screening a set of plasmids containing the mutant sequences against every assay on the panel. The CT values generated from these experiments are shown in Tables 2 and 3 and Table S2. A CT value of 30.0 indicates no amplification and that the specific mutation was not detected in that sample. Any CT value lower than 30.0 indicate amplification and those values generated by mutation-specific assays on their corresponding mutant plasmid are indicated in bold (Tables 2 and 3).
Table 2. Cross-reactivity matrix for the newly added assays in HRAS and PIK3CA.doi:10.1371/journal.pone.0090761.t002
Table 3. Cross-reactivity matrix for the newly added assays in HRAS and PIK3CA.doi:10.1371/journal.pone.0090761.t003
By utilizing the new AS-PCR assays, we were able to prevent the cross-reactivity found in certain instances on our previous panel (Tables 2 and 3). This highlights the specificity of our assays as some of the mutations are in the exact same position but have a single altered base, as in the case of Hr_G12S (position 34 G>A) and Hr_G12C (position 34 G>T) in Table 2.
The reproducibility of the mutation detection assays were evaluated by the comparison of duplicate experiments. The inter- and intra-chip variability in assay CT values was examined as shown in Figure 2. Inter-chip reproducibility was accessed by directly comparing the CT values of the mutant control between two chips and the Pearson correlation coefficient (r2) was calculated to be 0.995. A total of 5290 duplicate pairs were mapped on a scatter plot to determine the intra-chip reproducibility and the r2 value was found to be 0.990.
Figure 2. Quality control process for panel validation: Intra- and inter-chip reproducibility.
MUT-MAP panel qPCR assays were run in duplicate and CT outputs were plotted to determine both intra- and inter-chip reproducibility. Data for a typical mutation panel run are shown, with r2 values of 0.995 and 0.990 for inter- and intra-chip reproducibility, respectively.doi:10.1371/journal.pone.0090761.g002
To insure that no variability was introduced by different operator analysis, data from a single MUT-MAP experiment was analyzed by three independent operators. The CTs for the mutant control were found to have an r2 value of 0.993 after multiple regression analysis (data not shown).
Assay Sensitivity and Linearity
When sensitivity of assays were assessed by diluting plasmids serially either in nuclease-free water or a constant wild-type genomic DNA background (10 ng), most assays showed a lower limit of detection (LLOD) of 0.1–0.2% with a few exceptions. A few examples of such sensitivity analysis are shown in Figure 3 and the remaining data is shown in Figure S1. The wild-type and mutant CTs for these samples are graphed in blue, clearly showing that in the constant wild-type genomic DNA background the indicated mutation can be detected down to LLOD of 0.1–0.2% with a few exceptions as marked in Figure 3. The plasmid diluted in nuclease-free water (red squares) illustrates excellent linearity of the assays.
Figure 3. Evaluation of assay sensitivity.
Linearized plasmids containing the mutant sequence were mixed and diluted into a background of wild-type genomic DNA from 50-0.1% mutant (blue diamonds). A sample containing 5% of the corresponding mutant plasmid with a wild-type genomic DNA background was diluted in nuclease-free water (red squares). Samples were run on the panel and assay sensitivity was determined. The CT of wild-type genomic DNA alone is indicated by the green triangles.doi:10.1371/journal.pone.0090761.g003
Validation of Cell Line Samples
For cell line samples, gene-specific custom algorithms were written, taking into account the control CT and the mutant CT values. Samples showing ΔCT<6 were determined as positive for the specific mutation.
Over 600 cell lines have been analyzed by the MUT-MAP to detect mutations across the eleven genes. Table 4 highlights some of the cell lines that were found to have mutations that were detected by the newly added assays. These mutation calls were compared with the published characteristics of these cell lines annotated in the COSMIC database .
Table 4. Correlation Between Mutation Calls in Cell Lines and Those Reported in the Literature.doi:10.1371/journal.pone.0090761.t004
Benchmarking Sensitivity of MUT-MAP with NGS
To assess the accuracy and sensitivity of the MUT-MAP, we compared it with a commonly used NGS platform. Seven Latin Square mixes were formulated by mixing nine different FFPE samples containing twelve hotspot mutations (AKT1 E17K, BRAF V600E, EGFR deletion and L858R, HRAS Q61R, KRAS G12A, D, S and G13D, MET T1010I, and PIK3CA E545K and H1047L). When possible, the percentage of each mutation in the parental samples was determined by the SuraSeq500 panel (Figure 4A). Based on these percentages, the amount of each mutation in the seven Latin Square mixes were calculated and ranged from 0.14–32% (Figure 4B). By analyzing these samples on both platforms we were able to directly compare the sensitivity of twelve of our assays with the SuraSeq500 panel (Figure 4C).
Figure 4. Comparison of the sensitivity of MUT-MAP and a next generation sequencing platform.
(A and B) Nine FFPE samples with known mutation status were mixed together in varying concentrations following a Latin Square design to generate a seven-member Latin Square panel. The percentage of the mutant allele in each mix was calculated based on the mutant fraction of the parental samples as determined by analysis with the SuraSeq500 panel. For those mutations not detected by the NGS panel, 50% mutation in the parental sample was assumed. (C) The seven Latin Square samples were analyzed on MUT-MAP as well as by the SuraSeq500 panel on Ion Torrent in order to compare mutation calls and sensitivity levels of both platforms.doi:10.1371/journal.pone.0090761.g004
MUT-MAP was able to detect down to a 1.87% mutation for PIK3CA H1047X while NGS detected down to 0.94%. For BRAF V600E, MUT-MAP utilizes a TaqMan assay which was found to be less sensitive than the SuraSeq500 panel (9.05% and 0.28% respectively). Both platforms showed similar sensitivity to the AKT1 E17K mutation, as well as, the KRAS G12A and D, and G13D mutations. For the PIK3CA E545X and KRAS G12S mutations, both platforms were able to detect the lowest concentration present in our Latin Square mixes. The MUT-MAP panel also was able to detect HRAS Q61R down to a frequency of 0.39% while the SuraSeq500 panel did not detect the mutation at all in the Latin Square mixes or in the parental sample.
Disease-Specific Prevalence Study Analyses
We have performed oncogene mutation profiling on over 1000 individual tumor samples, including FFPE samples, from various cancer types. As an example, using the data generated with MUT-MAP we were able to determine the prevalence of specific mutations in breast and colon cancer (Figure 5A and B, respectively). For a collection of over 500 breast cancer samples we found 29.1% PIK3CA mutations, which is consistent with the COSMIC database , , , . We observed many KRAS (52.9%), PIK3CA (12.4%), and NRAS (7.4%) mutations in a colon cancer tissue collection (N = 121). The prevalence of these mutations also correlate well with those listed in the COSMIC database and other literature , , , , , , , . These results show that MUT-MAP is a sensitive and accurate platform to determine the mutational status in FFPE tissues and may be utilized to classify patients in clinical trials who may derive greater benefit with a targeted therapy.
Targeted therapies based on the mutational profiles of the tumor have become increasingly important in cancer diagnostics. We report here an updated MUT-MAP with expanded mutational coverage that includes 120 hotspot mutations in eleven cancer related genes. This panel requires as little as 2 ng of high quality gDNA from fresh frozen tissues or 100 ng of gDNA from FFPE tissues and validation using mutant plasmids showed robust assay signal and low cross-reactivity with all of the newly added assays. Mutation calls in cell lines were found to be consistent with the COSMIC database and MUT-MAP showed a 0.45% sensitivity in FFPE samples.
In comparison to the SuraSeq500 panel we have demonstrated that MUT-MAP is more sensitive in detecting the HRAS Q61R mutation in FFPE samples and has a similar sensitivity for detecting AKT1 E17K, KRAS G12A, D, and G13D mutations. SuraSeq500 was more sensitive in detecting BRAF V600E and EGFR L858R. Furthermore, MUT-MAP was able to detect these mutations with a much shorter turnaround time from start to finish, including data analysis, than the NGS platform used. While MUT-MAP lacks the breadth of coverage and flexibility of NGS, the platform can accurately and reliably detect hotspot mutations down to 0.45% (KRAS G12A) with very little FFPE DNA input. To date, we have utilized the platform to support multiple clinical programs and to study the prevalence of mutations in various disease settings to assist decision-making in drug development.
In conclusion, we describe here the development and validation of MUT-MAP, a high-sensitivity microfluidics chip-based mutation analysis panel to assay 120 hotspots across eleven oncogenes. This panel can rapidly and accurately determine the mutation status of cancer patient samples in a cost-effective and high-throughput manner. The mutation profiling data generated by MUT-MAP can be used to guide clinical decision-making and inform future clinical trial designs that could aid in the development of personalized health care.
Evaluation of assay sensitivity and linearity.
The preamplification primer sequences for the new MUT-MAP content: oncogenes PIK3CA, HRAS, FGFR3, FLT3, KIT and MET.
Cross-reactivity matrix for the newly added assays in FGFR3, FLT3, KIT, and MET.
We would like to thank Suresh Selvaraj for providing genomic DNA from all of the cell lines analyzed in the manuscript.
Conceived and designed the experiments: RR RP. Performed the experiments: RT E. Schleifman RP RD. Analyzed the data: RP E. Schleifman RT TS. Contributed reagents/materials/analysis tools: RP AT RT RD LF NS TM KB E. Smith RR. Wrote the paper: E. Schleifman RT RP AT NS TM KB E. Smith RR.
- 1. Gallo V, Egger M, McCormack V, Farmer PB, Ioannidis JP, et al. (2007) STrengthening the Reporting of OBservational studies in Epidemiology–Molecular Epidemiology (STROBE-ME): an extension of the STROBE Statement. PLoS Med 8: e1001117. doi: 10.1371/journal.pmed.1001117
- 2. Halaban R, Zhang W, Bacchiocchi A, Cheng E, Parisi F, et al. (2010) PLX4032, a selective BRAF(V600E) kinase inhibitor, activates the ERK pathway and enhances cell migration and proliferation of BRAF melanoma cells. Pigment Cell Melanoma Res 23: 190–200. doi: 10.1111/j.1755-148x.2010.00685.x
- 3. Mittempergher L, de Ronde JJ, Nieuwland M, Kerkhoven RM, Simon I, et al. (2011) Gene expression profiles from formalin fixed paraffin embedded breast cancer tissue are largely comparable to fresh frozen matched tissue. PLoS One 6: e17163. doi: 10.1371/journal.pone.0017163
- 4. Kompier LC, Lurkin I, van der Aa MN, van Rhijn BW, van der Kwast TH, et al. (2010) FGFR3, HRAS, KRAS, NRAS and PIK3CA mutations in bladder cancer and their potential as biomarkers for surveillance and therapy. PLoS One 5: e13821. doi: 10.1371/journal.pone.0013821
- 5. Zheng R, Bailey E, Nguyen B, Yang X, Piloto O, et al. (2011) Further activation of FLT3 mutants by FLT3 ligand. Oncogene 30: 4004–4014. doi: 10.1038/onc.2011.110
- 6. Beadling C, Jacobson-Dunlop E, Hodi FS, Le C, Warrick A, et al. (2008) KIT gene mutations and copy number in melanoma subtypes. Clin Cancer Res 14: 6821–6828. doi: 10.1158/1078-0432.ccr-08-0575
- 7. Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, et al. (1998) Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 279: 577–580. doi: 10.1126/science.279.5350.577
- 8. Schmidt L, Junker K, Weirich G, Glenn G, Choyke P, et al. (1998) Two North American families with hereditary papillary renal carcinoma and identical novel mutations in the MET proto-oncogene. Cancer Res 58: 1719–1722. doi: 10.1097/00005392-199905000-00137
- 9. Di Renzo MF, Olivero M, Martone T, Maffe A, Maggiora P, et al. (2000) Somatic mutations of the MET oncogene are selected during metastatic spread of human HNSC carcinomas. Oncogene 19: 1547–1555. doi: 10.1038/sj.onc.1203455
- 10. Ma PC, Kijima T, Maulik G, Fox EA, Sattler M, et al. (2003) c-MET mutational analysis in small cell lung cancer: novel juxtamembrane domain mutations regulating cytoskeletal functions. Cancer Res 63: 6272–6281. doi: 10.1016/s0169-5002(03)91964-7
- 11. Patel R, Tsan A, Tam R, Desai R, Schoenbrunner N, et al. (2012) Mutation scanning using MUT-MAP, a high-throughput, microfluidic chip-based, multi-analyte panel. PLoS One 7: e51153. doi: 10.1371/journal.pone.0051153
- 12. Reichert F BK, Myers TW, Schoenbrunner NJ, San Filippo J (2012) “DNA polymerases with increased 3′-mismatch discrimination”. US Patent 2011/0312041A1.
- 13. Will SG, Tsan A, Newton N (2011) “Allele-Specific Amplification,” US Patent 2010/0099110A1.
- 14. Hadd AG, Houghton J, Choudhary A, Sah S, Chen L, et al. (2013) Targeted, high-depth, next-generation sequencing of cancer genes in formalin-fixed, paraffin-embedded and fine-needle aspiration tumor specimens. J Mol Diagn 15: 234–247. doi: 10.1016/j.jmoldx.2012.11.006
- 15. Forbes SA, Bhamra G, Bamford S, Dawson E, Kok C, et al. (2008) The Catalogue of Somatic Mutations in Cancer (COSMIC). Curr Protoc Hum Genet Chapter 10 Unit 10 11. doi: 10.1002/0471142905.hg1011s57
- 16. O'Brien C, Wallin JJ, Sampath D, GuhaThakurta D, Savage H, et al. (2010) Predictive biomarkers of sensitivity to the phosphatidylinositol 3′ kinase inhibitor GDC-0941 in breast cancer preclinical models. Clin Cancer Res 16: 3670–3683. doi: 10.1158/1078-0432.ccr-09-2828
- 17. Saal LH, Holm K, Maurer M, Memeo L, Su T, et al. (2005) PIK3CA mutations correlate with hormone receptors, node metastasis, and ERBB2, and are mutually exclusive with PTEN loss in human breast carcinoma. Cancer Res 65: 2554–2559. doi: 10.1158/0008-5472-can-04-3913
- 18. Stemke-Hale K, Gonzalez-Angulo AM, Lluch A, Neve RM, Kuo WL, et al. (2008) An integrative genomic and proteomic analysis of PIK3CA, PTEN, and AKT mutations in breast cancer. Cancer Res 68: 6084–6091. doi: 10.1158/0008-5472.can-07-6854
- 19. Carpten JD, Faber AL, Horn C, Donoho GP, Briggs SL, et al. (2007) A transforming mutation in the pleckstrin homology domain of AKT1 in cancer. Nature 448: 439–444. doi: 10.1038/nature05933
- 20. Fumagalli D, Gavin PG, Taniyama Y, Kim SI, Choi HJ, et al. (2010) A rapid, sensitive, reproducible and cost-effective method for mutation profiling of colon cancer and metastatic lymph nodes. BMC Cancer 10: 101. doi: 10.1186/1471-2407-10-101
- 21. Kim MS, Jeong EG, Yoo NJ, Lee SH (2008) Mutational analysis of oncogenic AKT E17K mutation in common solid cancers and acute leukaemias. Br J Cancer 98: 1533–1535.
- 22. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, et al. (2010) Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol 11: 753–762. doi: 10.1016/s1470-2045(10)70130-3
- 23. Rizzo S, Bronte G, Fanale D, Corsini L, Silvestris N, et al. (2010) Prognostic vs predictive molecular biomarkers in colorectal cancer: is KRAS and BRAF wild type status required for anti-EGFR therapy? Cancer Treat Rev 36 Suppl 3: S56–61. doi: 10.1016/s0305-7372(10)70021-9
- 24. Tejpar S, Bertagnolli M, Bosman F, Lenz HJ, Garraway L, et al. (2010) Prognostic and predictive biomarkers in resected colon cancer: current status and future perspectives for integrating genomics into biomarker discovery. Oncologist 15: 390–404. doi: 10.1634/theoncologist.2009-0233
- 25. Samuels Y, Wang Z, Bardelli A, Silliman N, Ptak J, et al. (2004) High frequency of mutations of the PIK3CA gene in human cancers. Science 304: 554. doi: 10.1126/science.1096502
- 26. Irahara N, Baba Y, Nosho K, Shima K, Yan L, et al. (2010) NRAS mutations are rare in colorectal cancer. Diagn Mol Pathol 19: 157–163. doi: 10.1097/pdm.0b013e3181c93fd1
- 27. Janku F, Lee JJ, Tsimberidou AM, Hong DS, Naing A, et al. (2011) PIK3CA mutations frequently coexist with RAS and BRAF mutations in patients with advanced cancers. PLoS One 6: e22769. doi: 10.1371/journal.pone.0022769
- 28. Vaughn CP, Zobell SD, Furtado LV, Baker CL, Samowitz WS (2011) Frequency of KRAS, BRAF, and NRAS mutations in colorectal cancer. Genes Chromosomes Cancer 50: 307–312. doi: 10.1002/gcc.20854
- 29. Amado RG, Wolf M, Peeters M, Van Cutsem E, Siena S, et al. (2008) Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol 26: 1626–1634. doi: 10.1200/jco.2007.14.7116
- 30. Neumann J, Zeindl-Eberhart E, Kirchner T, Jung A (2009) Frequency and type of KRAS mutations in routine diagnostic analysis of metastatic colorectal cancer. Pathol Res Pract 205: 858–862. doi: 10.1016/j.prp.2009.07.010