Search tips
Search criteria 


Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS ONE. 2007; 2(5): e426.
Published online 2007 May 9. doi:  10.1371/journal.pone.0000426
PMCID: PMC1855985

Mutational Analysis of EGFR and Related Signaling Pathway Genes in Lung Adenocarcinomas Identifies a Novel Somatic Kinase Domain Mutation in FGFR4

Jörg Hoheisel, Academic Editor



Fifty percent of lung adenocarcinomas harbor somatic mutations in six genes that encode proteins in the EGFR signaling pathway, i.e., EGFR, HER2/ERBB2, HER4/ERBB4, PIK3CA, BRAF, and KRAS. We performed mutational profiling of a large cohort of lung adenocarcinomas to uncover other potential somatic mutations in genes of this signaling pathway that could contribute to lung tumorigenesis.

Methodology/Principal Findings

We analyzed genomic DNA from a total of 261 resected, clinically annotated non-small cell lung cancer (NSCLC) specimens. The coding sequences of 39 genes were screened for somatic mutations via high-throughput dideoxynucleotide sequencing of PCR-amplified gene products. Mutations were considered to be somatic only if they were found in an independent tumor-derived PCR product but not in matched normal tissue. Sequencing of 9MB of tumor sequence identified 239 putative genetic variants. We further examined 22 variants found in RAS family genes and 135 variants localized to exons encoding the kinase domain of respective proteins. We identified a total of 37 non-synonymous somatic mutations; 36 were found collectively in EGFR, KRAS, BRAF, and PIK3CA. One somatic mutation was a previously unreported mutation in the kinase domain (exon 16) of FGFR4 (Glu681Lys), identified in 1 of 158 tumors. The FGFR4 mutation is analogous to a reported tumor-specific somatic mutation in ERBB2 and is located in the same exon as a previously reported kinase domain mutation in FGFR4 (Pro712Thr) in a lung adenocarcinoma cell line.


This study is one of the first comprehensive mutational analyses of major genes in a specific signaling pathway in a sizeable cohort of lung adenocarcinomas. Our results suggest the majority of gain-of-function mutations within kinase genes in the EGFR signaling pathway have already been identified. Our findings also implicate FGFR4 in the pathogenesis of a subset of lung adenocarcinomas.


Lung cancer is the leading cause of cancer-related death in the United States and worldwide [1]. Despite recent advances in the treatment of lung cancer, the overall 5-year survival in the United States remains only 15%, highlighting the need for novel treatment strategies.

Lung cancers are currently classified into two major groups depending on histology: small cell lung cancer and non-small cell lung cancer (NSCLC). The latter is comprised of three different subtypes: adenocarcinoma, squamous cell carcinoma, and large cell carcinoma. The incidence of the adenocarcinoma subtype has been rising and now accounts for >50% of all cases of lung cancer [2]. Standard treatment for metastatic lung cancer involves empiric cytotoxic chemotherapy.

In order to develop specific therapies based upon the genetic makeup of individual NSCLC tumors, we (the Lung Cancer Oncogenome Group at Memorial Sloan-Kettering Cancer Center (MSKCC)) and others have sought to define clinically relevant molecular subsets of lung cancer. For example, we and others have shown that tumors highly sensitive to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (i.e. gefitinib or erlotinib) often contain dominant mutations in exons which encode a portion of the tyrosine kinase (TK) domain of EGFR [3][5]. Conversely, tumors with somatic mutations in KRAS, which encodes a GTPase downstream of EGFR, are resistant to therapy with these drugs [6][8]. Furthermore, about half of tumors with acquired resistance to these drugs display a second-site mutation in EGFR (Thr790Met) [9], [10]. Taken together, these data suggest that molecularly defined subgroups of lung cancer indeed exist and can be used to predict sensitivity and resistance to gefitinib and erlotinib. Clinicians in the future may be able to prescribe additional targeted therapies for patients with NSCLC based upon specific molecular characteristics.

At least six EGFR signaling pathway genes have been found to be mutated in NSCLC. While EGFR and KRAS mutations are detected in ~10% and 20% of NSCLCs, respectively, somatic mutations have also been identified in HER2/ERBB2 (~2%; exons 19 and 20) [11], [12] and HER4 (~2%, exons 20, 23) [13], the lipid kinase PIK3CA (~4%; exon 9) [14], and the serine/threonine kinase BRAF (~2%; exons 11 and 15) [15][17]. Most of these alterations have been found to be gain-of-function mutations. Except for PIK3CA mutations [18], [19], mutations in one of the other five genes are rarely found to be accompanied by a mutation in any of the remaining four, suggesting that they may have functionally equivalent roles in lung tumorigenesis [20]. All of these mutations are predominantly found in tumors with adenocarcinoma histology.

To uncover other potential gain-of-function somatic mutations that could have biological and clinical relevance in lung cancer, we performed mutational profiling of a large cohort of lung tumors, mostly adenocarcinomas. Because multiple genes that encode proteins in the EGFR signaling pathway have been found to be mutated in lung adenocarcinomas, we specifically sought to identify potential gain-of-function mutations in gene families in this pathway, i.e. in ERBB1-4, PIK3CA, AKT1-3, FRAP1, RPS6K1-2, RAS (K-, N-, and H-), RAF (A-, B-, C-), MAP2K1-2, and MAPK-1-3. We extended our studies to include other members of the MAP2K and MAPK gene families. We also examined FGFR1-4, because overexpression of FGF ligands in mouse lung epithelia leads to alveolar type II cell hyperplasia and adenomas [21][23]. All 39 genes have been reported to be expressed in mammalian lung tissues.


Tissue procurement

Resected tumor and matched normal adjacent lung specimens were obtained with patients' consent from the Memorial Sloan-Kettering Cancer Center (MSKCC) lung cancer tissue bank via a protocol approved by the Institutional Review Board (protocol #92-055). At the time of resection, samples were snap-frozen in the operating room in liquid nitrogen and then stored at minus 80°C until the time of use. Specimens were reviewed by a single pathologist (MFZ) for ≥70% tumor content and for histological verification. Clinical information was obtained from existing institutional databases. Some data regarding the mutation status of EGFR was previously reported [5].

Mutational profiling

Tumors selected for analyses were enriched for lung adenocarcinomas but were otherwise randomly selected, based upon availability of tissue. Squamous cell carcinomas were included to fill-in otherwise empty plate wells. No large cell carcinomas were studied.

DNA was extracted from tumors using a kit (DNeasy, Qiagen) or standard phenol extraction. Whole genome amplification (WGA) was performed by Qiagen. High-throughput (96-well plate) bidirectional dideoxynucleotide sequencing of PCR-amplified gene products was performed at the Genome Sequencing Center (Washington University in St. Louis) as per standard protocol ( The primer list can be found at: = 7.

PolyPhred [24] and PolyScan [25] software were used to generate an initial “automated” report of sequence variations. Tumor sequences were compared against reference sequences listed in the NCBI (RefSeq) database for each respective gene (see Supplemental Table S1). After visual inspection of the individual forward and reverse chromatograms for confirmation of non-synonymous sequence variations and insertions or deletions (including duplications), a “manual review” list of potential nucleotide changes was produced. Synonymous variants and those with corresponding dbSNP ( entries were also excluded.

Mutation verification

Putative kinase domain mutations listed in the manual report were subsequently verified at MSKCC by bidirectional sequence analysis of a separate individual PCR product. Variants were deemed somatic if they were found to be absent in matched normal tissue. Primers were designed to detect each individual mutation, using each respective reference sequence and Vector NTI (Supplemental Tables S1 and S2). All PCR reactions were performed with HotStarTaq Master Mix Kit (Qiagen, Valencia, California), using standard conditions (95°C×15 min; 95°C×30 s, 60°C×30 s, 72°C×60 s, for 36 cycles, then 72°C for 5 minutes, 50 µl reactions). PCR products were purified with a MultiScreen Resist vacuum manifold and PCR96 Cleanup Plates (Millipore). Sequencing reactions were performed using Applied Biosystems Version 3.1 Big Dye Terminator chemistry and analyzed on an Applied Biosystems 3730 Sequencer.

Development of “Mutagrator” – a mutation interpretation tool for tyrosine kinases

To support the interpretation of putative kinase domain mutations, we created a prototype mutation interpretation tool for tyrosine kinases (TKs), called “Mutagrator”, located at (freely available to the research community). Mutagrator is a software program which takes curated mutation data from the literature and displays it in the context of a master protein (chosen by the user) and a protein-registered TK multiple domain alignment. In order to create the multiple alignment, we first retrieved 108 human TK gene records from EntrezGene by querying for domain cd00192 [26]. We then extracted TK domains from all 168 protein isoforms corresponding to these genes from Entrez Protein [27], aligned the domains using the ClustalW program [28], and added additional feature information, including ATP binding residues, activation loop, catalytic loop and substrate binding site boundaries from Conserved Domain Database (CDD) [29]. All input and output files are available on the Mutagrator website. Currently, curated mutation data is ingested from the Catalogue of Somatic Mutations in Cancer (COSMIC), which was created and is maintained by the Sanger Institute [30]. Collected data includes mutation (amino acid change and position), mutation type (point, insertion, deletion, complex), involved gene, tissue type, cancer type and published source. The version of the database used in this study (v20) consisted of about 30,000 individual mutations in about 1,300 genes, and corresponding to about 3,300 distinct mutations. From these data, Mutagrator produced interlinked, static HTML webpages of two types: master protein pages (for each protein in the TK domain alignment with mutations), and detailed mutation pages (for each protein residue position).


We screened coding sequences from 39 genes for mutations in genomic DNA from a total of 261 resected, clinically annotated non-small cell lung cancer (NSCLC) specimens. 90% of tumors were adenocarcinomas, and 10% were squamous cell carcinomas. Clinical characteristics of examined tumors are listed in Supplemental Table S3, and the exonic coverage of genes is listed in Supplemental Table S1.

Due to logistical reasons, the mutational analysis was performed in two partially overlapping groups. We first examined genomic DNAs from 217 tumors for mutations in a set of core genes previously reported to harbor mutations in NSCLC, i.e. in EGFR, HER2, HER4, KRAS, PIK3CA, and BRAF (Figure 1). We also profiled HER3, MAP2K4, and FGFR1-4 (Figure 1). We then examined 93 WGA-treated DNA tumor samples for mutations in EGFR pathway genes and a set of exploratory genes (Figure 1). Ten genes were sequenced in both groups (Figure 1) to maximize the number of tumors sequenced for the core genes. Eighty percent of the sequence reads in the WGA-treated specimens had a Phred quality score of at least 20 (data not shown), suggesting that most base-calling had an accuracy of 99% [31].

Figure 1
Genes sequenced in this study.

Automatic and manual sequence analyses (see methods) identified 239 putative non-synonymous sequence variations, comprised of 174 different types of variants that differed from published sequences (Figure 2, and Supplemental Table S1). To focus our efforts, we concentrated on further examining the 22 variants (6 types) found in 3 RAS family genes and the 135 variants (99 types) found within exons encoding kinase domains of kinases. The 82 non-kinase domain variants (69 distinct types) have not yet been examined, although none occur at a frequency higher than 2%.

Figure 2
Schematic of overall results.

We confirmed 21 of the sequence variations in the RAS family. 20 were somatic (all in codons 12 or 13 of exon 2 of KRAS), while one in HRAS was found in matched normal DNA (Supplemental Table S4). The prevalence of KRAS mutations in our cohort of lung adenocarcinomas was 12% (20/173). All confirmed somatic mutations were found in adenocarcinomas except for a Gly12Asp mutation in KRAS in a squamous cell carcinoma (Table 1).

Table 1
Clinical characteristics of patients whose tumors contained a somatic mutation.

67 of the 135 kinase domain sequence variations were confirmed by analysis of sequence tracings from an independent PCR isolate. 48 variants were also found in corresponding normal samples (Supplemental Table S4). Two were of uncertain significance, because we were unable to obtain a PCR product from DNA from matched normal tissue (Supplemental Table S4). Of the remaining 17 confirmed non-synonymous somatic variants, 16 were found in genes known to be mutated in NSCLC, i.e. EGFR, BRAF, and PIK3CA (Supplemental Table S4). The prevalence of EGFR, BRAF, and PIK3CA mutations in lung adenocarcinomas was 6 (13/234), <1 (1/156), and 2% (2/132), respectively. Clinical characteristics of all tumors containing somatic mutations can be found in Table 1. One PIK3CA mutation was found in a tumor that also contained a KRAS mutation. No other tumor had more than one somatic mutation (Table 2).

Table 2
Mutations observed in EGFR, KRAS, BRAF, and FGFR4 in lung adenocarcinomas.

In one lung adenocarcinoma specimen from a current smoker, we found a somatic heterozygous G to A mutation at nucleotide position 2041 in exon 16 of FGFR4 (Figure 3). This mutation would lead to substitution of lysine for glutamic acid at position 681 (Glu681Lys), 51 amino acids downstream of the highly conserved DFG motif found in all protein kinases (Figure 4). Using our “Mutagrator tool” (Figure 4; see methods), we determined that an analogous mutation has been reported in a glioblastoma in ERBB2 (Glu914Lys) [11]. Moreover, the glutamic acid at position 681 is highly conserved among various kinases (Figure 4). The biological significance of the lung FGFR4 mutation remains to be determined experimentally. In total, this mutation was found in 1 of 158 tumors. We did not identify any other somatic mutations in this tumor (Table 2).

Figure 3
Analysis of FGFR4.
Figure 4
Amino acid alignment of the FGFR4 kinase domain with other tyrosine kinase domains found to be altered in human cancers.


We report a comprehensive sequencing study of major genes in a specific signaling pathway in a sizeable cohort of lung adenocarcinoma tumor specimens. Previous large-scale mutational profiling studies of lung cancer have examined either only the exons encoding the activation loops of receptor tyrosine kinase (RTK) genes (47 of 58 RTK genes) in 119 primary NSCLCs, of which 70 (59%) were lung adenocarcinomas [4], or the coding sequences of 518 protein kinases in a relatively limited number of samples, i.e. 26 primary lung neoplasms (7 adenocarcinomas) and seven cancer cell lines (6 adenocarcinomas) [32]. Here, we examined a total of 261 tumor samples, predominantly adenocarcinomas, specifically for genetic alterations in genes encoding major signaling proteins in the EGFR signaling pathway. We also determined the status of a select set of other genes potentially relevant to lung tumorigenesis.

Most of the somatic mutations we found have been reported, including mutations in EGFR, KRAS, BRAF, and PIK3CA [3][5], [11][17]. The relative distribution of these mutations in our lung adenocarcinomas matches that observed by others. The frequency of EGFR and KRAS mutations was slightly lower than other published series, possibly because the mutation detection software that we used went through various stages of development during this project [25]. We did not identify any somatic mutations in HER2 or HER4. However, one of two variants of uncertain significance (due to inability to PCR amplify a gene product from matched normal DNA) was located in the kinase domain of HER2 (Arg784Cys) and has not been previously reported.

We did find a novel mutation (Glu681Lys) in the kinase domain (exon 16) of FGFR4 in 1 of 158 tumors. This mutation is analogous to the previously reported Glu914Lys kinase domain mutation in ERBB2 found in a glioblastoma [11]. Glu681 is highly conserved region among various kinases, downstream of the DFG motif. Based on the crystal structure of the related family member FGFR1 tyrosine kinase domain (PDB accession 1FGK) [33], the analogous residue (Glu692) appears in close proximity to Ala626 in the TK catalytic loop and Arg661 in the TK activation loop. Since Glu692 is strongly positively charged and Arg661 is strongly negatively charged, the close spatial proximity of these two residues would likely lead to a strong ionic bond and therefore may be functionally important. Extrapolating back to FGFR4, we propose that the Glu681Lys mutation may alter the functional properties of the TK catalytic domain by reversing the charge of residue 681, potentially disrupting an ionic bond with residue Arg650, and thereby disrupting normal function of FGFR4 (Figure 5).

Figure 5
Structural modeling of the FGFR4 Glu681Lys amino acid substitution.

FGFR4 is a monomeric receptor protein tyrosine kinase possessing three immunoglobulin-like domains in the extracellular region. The protein is one of four high-affinity receptors for multiple members of the FGF family of ligands that evoke angiogenic, mitogenic, and differentiation responses in cells [34]. Such ligands, when overexpressed in mouse lung epithelia, stimulate alveolar type II cell hyperplasia and adenoma formation [21][23]. Interestingly, Davies et al have reported that a lung adenocarcinoma cell line also harbors a non-synonymous mutation in exon 16 of FGFR4 – Pro672Thr [32]. [The Davies et al paper referenced FGFR4 transcript variant 2; we referenced variant 1, so the equivalent mutation would be Pro712Thr.] Collectively, these data suggest a role for FGFR4 mutations in a subset of lung adenocarcinomas. The Sanger group also found two other somatic mutations in genes that encode the related family members, FGFR1 and FGFR2, in lung cancer specimens. The described FGFR1 and FGFR2 mutations occur outside the kinase domain, but in identical positions to activating germline mutations known to predispose to skeletal dysplasias. Other FGFR gene alterations have also been reported in human cancers, although rarely in exons encoding the kinase domain (reviewed in [34]). We plan to characterize the functional consequences of the two reported FGFR4 mutations and determine their prevalence in independent lung and other tumor specimen banks.

This study has some potential limitations. First, we examined only 39 genes. We did not sequence all related gene family members such as RPS6KA1-6, MAP2K3, and MAP2K7. This study also did not seek potential mutations in genes encoding adaptor proteins or phosphatases that might affect the ERBB signaling pathway. Second, WGA could have skewed the results by selectively amplifying DNA from normal rather than tumor tissue. However, evaluation of data from multiple assays has established that base-calling discrepancies between amplified and unamplified samples are minimal and not significantly different than that observed after re-sequencing non-amplified samples [35], [36]. Consistent with this, in all cases where we found an EGFR or KRAS mutation in the original non-WGA-treated sample, we also detected the same mutation in the corresponding WGA-treated sample (n = 14; data not shown). Finally, in this initial study, we restricted our verification studies to non-synonymous variants in the exons encoding kinase domains, in view of the clinical significance of known somatic mutations in kinase domains. The 69 types of non-kinase domain sequence variations we identified are currently undergoing confirmation. Nevertheless, the prevalence thus far of non-synonymous somatic mutations per megabase of tumor sequenced in this study was 4.1 (37 total mutations/9Mb). This rate is slightly higher than that found by others in a mutational analysis of ~13,000 genes in 11 colorectal and 11 breast cancers [37].

This study represents an early step towards an understanding of the lung cancer oncogenome. Our results suggest that the majority of gain-of-function mutations within kinase genes in the EGFR signaling pathway may have been identified. We await results from the NCI/NHGRI-sponsored “technical demonstration project” – a pilot project for The Cancer Genome Atlas initiative, in which approximately 200 highly-curated lung adenocarcinomas are being analyzed for chromosomal gains and losses simultaneously with mutational profiling of about 1000 genes thought to be relevant to lung tumorigenesis. Efforts such as these should contribute towards the identification of the full spectrum of somatic mutations found in lung adenocarcinomas.

Supporting Information

Table S1

Gene, GenBank accession number, and exonic coverage of genes sequenced in this study.

(0.06 MB DOC)

Table S2

List of primers used to verify putative variants.

(0.15 MB DOC)

Table S3

Clinical characteristics of patients whose tumors were analyzed. Group 1 was used for sequencing the “core” genes. Group 2 was used for sequencing the “exploratory” genes. Some tumors and genes overlapped between the two groups. Smoking history is defined as never smokers (<100 lifetime cigarettes), former smokers (quit ≥1 year prior to diagnosis), or current (quit <1 year prior to diagnosis). See text and Figure 1 for more detail. 1Adeno includes adenocarcinoma with bronchioalveolar features (n = 79, n = 27 for Group 1 and 2, respectively). Abbreviations: Adeno, adenocarcinoma; SCC, squamous cell carcinoma.

(0.04 MB DOC)

Table S4

List of variants verified. Group headings correspond to groups in bottom row of Figure 2. Variants found in normal tissue did not have an existing entry in dbSNP. 1A total of 5 EGFR mutations (exon 19 del, n = 1: exon 21 L858R, n = 4) have been previously reported (5). 2Variants with high frequency were not verified in all samples. If a variant was also found in DNA from five matched normals, no further samples were verified. Abbreviations: del, deletion; dup, duplication.

(0.08 MB DOC)


We thank the Lung Cancer Oncogenome Group at MSKCC for helpful discussions; Katerina Politi and Charles Sawyers for critical reading of the manuscript; and Dhananjay Chitale and Marc Ladanyi for help with DNA specimen extraction.


Competing Interests: The authors have declared that no competing interests exist.

Funding: This work was supported by the Doris Duke Charitable Foundation (WP), Joan's Legacy: The Joan Scarangello Foundation to Conquer Lung Cancer (WP), NIH K08-CA097980 (WP), the Bristol-Myers Squibb Foundation (AL), the Jodi Spiegel Fisher Cancer Foundation (WP), and The Carmel Hill Fund. WP is the incumbent of the Miner Family Chair for Intrathoracic Cancers. The funders of this study had no role in the design and conduct of the study, in the collection, analysis and interpretation of the data, or preparation and review of the manuscript.


1. Jemal A, Siegel R, Ward E, Murray T, Xu J, et al. Cancer statistics, 2006. CA Cancer J Clin. 2006;56:106–130. [PubMed]
2. Gabrielson E. Worldwide trends in lung cancer pathology. Respirology. 2006;11:533–538. [PubMed]
3. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med. 2004;350:2129–2139. [PubMed]
4. Paez JG, Janne PA, Lee JC, Tracy S, Greulich H, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science. 2004;304:1497–1500. [PubMed]
5. Pao W, Miller V, Zakowski M, Doherty J, Politi K, et al. EGF receptor gene mutations are common in lung cancers from “never smokers” and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci USA. 2004;101:13306–13311. [PubMed]
6. Pao W, Wang TY, Riely GJ, Miller VA, Pan Q, et al. KRAS mutations and primary resistance of lung adenocarcinomas to gefitinib or erlotinib. PLoS Medicine. 2005;2:e17. [PMC free article] [PubMed]
7. Han S-W, Kim T-Y, Hwang PG, Jeong S, Kim J, et al. Predictive and prognostic impact of epidermal growth factor receptor mutation in non-small-cell lung cancer patients treated with gefitinib. J Clin Oncol. 2005;23 [PubMed]
8. Giaccone G, Gallegos Ruiz M, Le Chevalier T, Thatcher N, Smit E, et al. Erlotinib for frontline treatment of advanced non-small cell lung cancer: a phase II study. Clin Cancer Res. 2006;12:6049–6055. [PubMed]
9. Pao W, Miller VA, Politi KA, Riely GJ, Somwar R, et al. Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib is associated with a second mutation in the EGFR kinase domain. PLoS Medicine. 2005;2:e73. [PMC free article] [PubMed]
10. Kobayashi S, Boggon TJ, Dayaram T, Janne PA, Kocher O, et al. EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. New Engl J Med. 2005;352:786–792. [PubMed]
11. Stephens P, Hunter C, Bignell G, Edkins S, Davies H, et al. Lung cancer: intragenic ERBB2 kinase mutations in tumours. Nature. 2004;431:525–526. [PubMed]
12. Shigematsu H, Takahashi T, Nomura M, Majmudar K, Suzuki M, et al. Somatic mutations of the HER2 kinase domain in lung adenocarcinomas. Cancer Res. 2005;65:1642–1646. [PubMed]
13. Soung YH, Lee JW, Kim SY, Wang YP, Jo KH, et al. Somatic mutations of the ERBB4 kinase domain in human cancers. Int J Cancer. 2006;118:1426–1429. [PubMed]
14. Samuels Y, Wang Z, Bardelli A, Silliman N, Ptak J, et al. High frequency of mutations of the PIK3CA gene in human cancers. Science. 2004;304:554. [PubMed]
15. Naoki K, Chen TH, Richards WG, Sugarbaker DJ, Meyerson M. Missense mutations of the BRAF gene in human lung adenocarcinoma. Cancer Res. 2002;62:7001–7003. [PubMed]
16. Brose MS, Volpe P, Feldman M, Kumar M, Rishi I, et al. BRAF and RAS mutations in human lung cancer and melanoma. Cancer Res. 2002;62:6997–7000. [PubMed]
17. Davies H, Bignell G, Cox C, Stephens P, Edkins S, et al. Mutations of the BRAF gene in human cancer. Nature. 2002;417:949–954. [PubMed]
18. Endoh H, Yatabe Y, Kosaka T, Kuwano H, Mitsudomi T. PTEN and PIK3CA expression is associated with prolonged survival after gefitinib treatment in EGFR-mutated lung cancer patients. J Thor Oncol. 2006;1:629–634. [PubMed]
19. Kawano O, Sasaki H, Endo K, Suzuki E, Haneda H, et al. PIK3CA mutation status in Japanese lung cancer patients. Lung Cancer. 2006;54:209–215. [PubMed]
20. Shigematsu H, Nomura M, Suzuki M, Wistuba II, Fujisawa T, et al. Gene mutation differences in lung cancers arising in never and ever smokers. Proc AACR. 2005;LB-39
21. Clark J, Tichelaar J, Wert S, Itoh N, Perl A, et al. FGF-10 disrupts lung morphogenesis and causes pulmonary adenomas in vivo. Am J Physiol Cell Mol Physiol. 2001;280:L705–L715. [PubMed]
22. Tichelaar J, Lu W, Whitsett J. Conditional expression of fibroblast growth factor-7 in the developing and mature lung. J Biol Chem. 2000;275:11858–11864. [PubMed]
23. Zhao B, Chua S, Burcin M, Reynolds S, Stripp B, et al. Phenotypic consequences of lung-specific inducible expression of FGF-3. Proc Natl Acad Sci U S A. 2001;98:5898–5903. [PubMed]
24. Nickerson DA, Tobe VO, Taylor SL. PolyPhred: automating the detection and genotyping of single nucleotide substitutions using fluorescence-based resequencing. Nucleic Acids Research. 1997;25:2745–2751. [PMC free article] [PubMed]
25. Chen K, McLellan MD, Ding L, Wendl MC, Kasai Y, et al. PolyScan: an automatic indel and SNP detection approach to the analysis of human re-sequencing data. Genome Res in press [PubMed]
26. Maglott D, Ostell J, Pruitt KD, Tatusova T. Entrez Gene: gene-centered information at NCBI. Nucleic Acids Res 2006 [PMC free article] [PubMed]
27. Wheeler DL, Barrett T, Benson DA, Bryant SH, Canese K, et al. Database resources of the National Center for Biotechnology Information. Nucleic Acids Res. 2006;34:D173–180. [PMC free article] [PubMed]
28. Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, et al. Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 2003;31:3497–3500. [PMC free article] [PubMed]
29. Marchler-Bauer A, Anderson JB, Derbyshire MK, Deweese-Scott C, Gonzales NR, et al. CDD: a conserved domain database for interactive domain family analysis. Nucleic Acids Res 2006 [PMC free article] [PubMed]
30. Forbes S, Clements J, Dawson E, Bamford S, Webb T, et al. Cosmic 2005. Br J Cancer. 2006;94:318–322. [PMC free article] [PubMed]
31. Ewing B, Hillier L, Wendl MC, Green P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Research. 1998;8:175–185. [PubMed]
32. Davies H, Hunter C, Smith R, Stephens P, Greenman C, et al. Somatic mutations of the protein kinase gene family in human lung cancer. Cancer Res. 2005;65:7591–7595. [PubMed]
33. Mohammadi M, Schlessinger J, Hubbard SR. Structure of the FGF receptor tyrosine kinase domain reveals a novel autoinhibitory mechanism. Cell. 1996;86:577–587. [PubMed]
34. Eswarakumar VP, Lax I, Schlessinger J. Cellular signaling by fibroblast growth factor receptors. Cytokine Growth Factor Rev. 2005;16:139–149. [PubMed]
35. Bredel M, Bredel C, Juric D, Kim Y, Vogel H, et al. Amplification of whole tumor genomes and gene-by-gene mapping of genomic aberrations from limited sources of fresh-frozen and paraffin-embedded DNA. J Mol Diagn. 2005;7:171–182. [PubMed]
36. Paez JG, Lin M, Beroukhim R, Lee JC, Zhao X, et al. Genome coverage and sequence fidelity of phi29 polymerase-based multiple strand displacement whole genome amplification. Nucleic Acids Research. 2004;32:e71. [PMC free article] [PubMed]
37. Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, et al. The Consensus Coding Sequences of Human Breast and Colorectal Cancers. Science. 2006;314:268–74. [PubMed]
38. Ebright MI, Zakowski MF, Martin J, Venkatraman ES, Miller VA, et al. Clinical pattern and pathologic stage but not histologic features predict outcome for bronchioloalveolar carcinoma. Ann Thorac Surg. 2002;74:1640–1646. [PubMed]
39. Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The new World Health Organization classification of lung tumours. Eur Respir J. 2001;18:1059–1068. [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science