IJC International Journal of Cancer
Genes involved in the WNT and vesicular trafficking pathways are associated with melanoma predisposition Maider Ibarrola-Villava1, Rajiv Kumar2, Eduardo Nagore3, Meriem Benfodda4,5, Mickael Guedj6, Steven Gazal7, Hui-Han Hu4,5, Jian Guan2, P. Sivaramakishna Rachkonda2, Vincent Descamps4,8, Nicole Basset-Seguin4,8, Armand Bensussan4, Martine Bagot4,9, Philippe Saiag10, Dirk Schadendorf11, Manuel Martin-Gonzalez12, Matias Mayor13, Bernard Grandchamp5, Gloria Ribas1 and Nadem Soufir4,5 1
Department of Haematology and Medical Oncology, Biomedical Research Institute INCLIVA, Valencia 46010, Spain Division of Molecular Genetic Epidemiology, German Cancer Research Center, Heidelberg 69120, Germany 3 Department of Dermatology, Instituto Valenciano de Oncologia, Valencia 46009, Spain 4 Inserm U976, Centre de Recherche Sur la Peau, Hopital Saint Louis, Universite Paris 7, Paris 75010, France 5 ^pital Bichat, APHP, Paris 75018, France Departement de Genetique, Ho 6 Laboratoire Statistiques et Genomes, Evry 91000, France 7 UMR S738, Faculte de Medecine Xavier Bichat, Paris 75018, France 8 Department of Dermatology, Hopital Bichat, APHP, Paris 75018, France 9 Department of Dermatology, Hopital Saint Louis, APHP, Paris 75010, France 10 Department of Dermatology, Hopital Ambroise Pare, APHP, Paris 92100, France 11 Department of Dermatology, University Hospital Essen, Esse 45147, Germany 12 Department of Dermatology, Hospital Ramon y Cajal, Madrid 28034, Spain 13 Department of Dermatology, Hospital La Paz, Madrid 28046, Spain
Multifactorial predisposition to melanoma includes genes involved in pigmentation, immunity and DNA repair. Nonetheless, missing heritability in melanoma is still important. We studied the role of 335 candidate SNPs in melanoma susceptibility by using a dedicated chip and investigating 110 genes involved in different pathways. A discovery set was comprised of 1069 melanoma patients and 925 controls from France. Data were replicated using validation phases II (1085 cases and 801 controls from Spain) and III (1808 cases and 1894 controls from Germany and a second set of Spanish samples). In addition, an exome sequencing study was performed in three high-risk French melanoma families. Nineteen SNPs in 17 genes were initially associated with melanoma in the French population. Six SNPs were replicated in phase II, including two new SNPs in the WNT3 (rs199524) and VPS41 (rs11773094) genes. The role of VPS41 and WNT3 was confirmed in a meta-analysis (3940 melanoma cases and 3620 controls) with two-side p values of 0.002, (OR 5 0.86) and 4.07 3 10210 (OR 5 0.80), respectively. Exome sequencing revealed a non-synonymous VPS41 variant in one family that was shown to be strongly associated with familial melanoma (OR 5 4.46, p 5 0.001) in an independent sample of 178 melanoma families. WNT3 belongs to WNT pathway known to play a crucial role in melanoma, whereas VPS41 regulates vesicular trafficking and is thought to play a role in pigmentation. Our work identified two new pathways involved in melanoma predisposition. These results may be useful in the future for identifying individuals highly predisposed to melanoma.
The knowledge of the genetic predisposition to melanoma has significantly increased during the last several years. Linkage studies, candidate gene approaches, genome-wide association studies (GWAS) and whole-exome sequencing studies have
identified several loci that influence susceptibility to cutaneous malignant melanoma (MM).1–3 Mutations in the highly penetrant CDKN2A (MIM#60160), CDK4 (MIM#609048), TERT (MIM#187270) and POT1 (MIM#606478) genes are well
Key words: exome sequencing, case-control study, WNT3, VPS41, melanoma Abbreviations: CI: confidence interval; GWAS: genome-wide association studies; HWE: Hardy Weinberg equilibrium; MAF: minor allele frequency; MM: malignant melanoma; OR: odds ratio; SNP: single-nucleotide polymorphism Additional Supporting Information may be found in the online version of this article. Grant sponsor: Ministerio de Salud Carlos III; Grant number: PI10-0405, CP08-00069; Grant sponsor: La Societe Francaise de Dermatologie (SFD), FEDER DOI: 10.1002/ijc.29257 History: Received 7 Apr 2014; Accepted 9 Sep 2014; Online 10 Oct 2014 Correspondence to: Nadem Soufir, Departement de Genetique, H^opital Bichat Claude Bernard, 46 rue Henri Huchard, 75018 Paris, France. Tel.: [33140258551], Fax: 1[33140258785], E-mail: nadem.soufi[email protected]
C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
Cancer Genetics
2
2
VPS41 and WNT3 and melanoma susceptibility
Cancer Genetics
What’s new? While our understanding of the molecular pathways that lead to malignant melanoma has increased over the past few years, it is still incomplete. In this study, the authors identified two new pathways: one that involves WNT signaling (involved in melanoblast development), and another that involves vesicular trafficking (involved in changes in pigmentation phenotype). These results may be useful in developing screening tests for people who are genetically susceptible to melanoma. In addition, the whole-exome sequencing techniques used in this study may aid researchers in identifying mutations in unknown diseasecandidate genes.
known to play a role in familial melanoma, which accounts for about 10% of all MM cases.4–6 Moreover, low penetrance genetic variants have been associated with sporadic form of the disease and, therefore, pigmentation, nevus count, immune responses, DNA repair, oxidative stress and the vitamin D receptor have been associated with melanoma susceptibility.7–12 Although different genetic pathways are involved in melanoma onset, the human pigmentation pathway seems to play a critical role in the pathogenesis of the disease. MM predisposition is increased in individuals characterized by fair skin colour, high mole count, blond or red hair and high sun sensibility (Fitzpatrick’s phototypes I and II).4 Since the incidence of malignant melanoma in part reflects the consequences of a complex interaction between skin phenotype and UV exposure, variations in genes implicated in pigmentation and tanning response to UV are associated with susceptibility to MM. Thus, comprehensive knowledge of the entire network of genetic variants in pigmentation and melanogenesis genes may be a useful predictor of MM risk in the general population.13,14 The pigmentation pathway is a multistep process that involved differentiation and migration of melanocytes (pigment-producing cells), melanosome biogenesis, and migration and melanin synthesis. More than 120 genes have been identified through animal models. Nevertheless, only one gene, the melanocortin 1 receptor (MC1R, MIM#155555) is known to account for a substantial variation in the incidence of sporadic MM risk.15 Several other genes, such as OCA2 (MIM#611409), ASIP (MIM#600201), SLC45A2 (MIM#606202), TYR (MIM#606933) or TYRP1 (MIM#115501) have also emerged as being potentially important in melanoma susceptibility.1, 16–18 Considering the involvement of pigmentation in MM predisposition, the goal of the current study was to identify novel variants associated with MM risk by using singlenucleotide polymorphism (SNP) array on three independent sporadic melanoma series of European samples: France, Germany and Spain. Moreover, we performed whole-exome sequencing on a series of three melanoma families.
Material and Methods Study subjects and data collection Discovery phase I. A total number of 1069 melanoma cases
were recruited from 2002 to 2008 at the Departments of Dermatology of five different university-affiliated hospitals from
the Paris region (Saint Louis Hospital, Bichat Hospital, Tarnier Hospital, Ambroise Pare Hospital and Tenon Hospital). This cohort, called MelanCohort, comprised 888 sporadic melanoma cases (83%), 61 multiple melanoma patients (5.7%) and 120 nonrelated familial melanoma cases (11.3%). Similarly, 925 skin cancer-free control subjects were recruited from the EFS (Etablissement Franc¸ais du Sang) at Bichat and Saint Louis hospitals (Paris region) during the same period. A standardised questionnaire was used to collect information on pigmentation characteristics, the presence of childhood sunburns, Fitzpatrick’s skin type classification and personal and family history of cancer. Validation phase II. A second phase of the study was com-
posed of 1085 nonrelated MM cases obtained from the Departments of Dermatology of three different hospitals in Madrid, Spain (Gregorio Mara~ non Hospital, La Paz Hospital and Ramon y Cajal Hospital) and from the Instituto Valenciano de Oncologia in Valencia, Spain. A total of 801 cancerfree controls from the college of lawyers (Madrid, Spain), Gregorio Mara~ non Hospital and Instituto Valenciano de Oncologia were recruited. All subjects came from the geographical regions covered by the hospitals involved in the study. Validation phase III. A third phase of the study was com-
posed of two different sample sets. The first sample set included 872 nonrelated MM cases and 1158 cancer-free controls obtained from a different collection from the Instituto Valenciano de Oncologia. The second one included 936 MM patients and 736 controls recruited from the German Cancer Research Centre (Heidelberg, Germany). A total of 1808 melanoma cases and 1894 control subjects were included in this phase III validation. All subjects came from the geographical regions covered by their respective study hospitals. All enrolled patients and controls were from Caucasian origin. Distribution of cases and controls in discovery and validation phases is summarised in Supporting Information Table 1 and Supporting Information Table 2. Genomic DNA from cases and controls was isolated from peripheral blood lymphocytes using the MagNA Pure LC Instrument according to the manufacturer’s protocol (Roche Applied Science, Mannheim, Germany). DNA concentration was quantified using Quant-iT PicoGreen dsDNA Reagent (Invitrogen, Eugene, OR). All DNA samples were subjected to a quality control before processing them on the Illumina C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
3
Ibarrola-Villava et al.
arrays (Illumina, San Diego, CA). All subjects gave written informed consent, and the study was approved by the Ethics Committees of the Gregorio Mara~ non Hospital, Hospital Clinico Universitario of Valencia and the Ambroise Pare Hospital (Boulogne, France).
quality control measure, one sample duplicate and a nontemplate sample per 96-well plate were included. For the first validation set (Phase II), a total of 13 SNPs were genotyped while 7 SNPs were studied in the second validation sample set (Phase III). Genotypes were scored by the LGC Genomics Company.
Gene and SNP selection
Exome sequencing
Previous literature and the information available on public databases were used for gene selection. Eventually, 110 genes were included in the study. The most overrepresented category with 96 genes was pigmentation. Pigmentation genes were chosen from the albinism and pigmentation genes databases (http://www.espcr.org/micemut/). A further group of 14 genes implicated in DNA repair (3 genes), cell cycle (5 genes), immune responses and apoptosis (4), MAP kinase pathway (1 gene) and glutamate receptor (1 gene) were included in the study. Three databases were used to collect information about single nucleotide polymorphisms (SNPs): NCBI (www.ncbi. nlm.nih.gov/SNP), HapMap (www.hapmap.org) and Illumina databases (www.illumina.com). Finally, 384 SNPs were chosen that included (a) polymorphisms predicted to have a functional impact (snpinfo.niehs.nih.gov/snpinfo/snpfunc.htm) that comprise SNPs located in putative transcription factor binding sites (66 SNPs) or in conserved regions (81 SNPs), nonsynonymous polymorphisms (49 SNPs), SNPs predicted to affect splicing (65 SNPs) or miRNA binding sites (63 SNPs), and (b) some tag-SNPs (87 SNPs). Tag-SNPs range from the hypothetical promoter area (2000 Kb upstream) until 2000 Kb downstream of the gene. Eight SNPs commonly associated with MM were included as positive controls (rs4911414 in ASIP, rs11547464, rs1805007 and rs1805009 in MC1R, rs16891982 in SLC45A2, rs1126809 in TYR, rs2733832 in TYRP1 and rs910873 in the 20Q region). To ensure a high genotyping success rate, a minor allele frequency threshold of 0.05 in the European HapMap population and a recommended manufacturer “Illumina score” were established in the SNP selection. SIFT and PolyPhen, (http:// sift.jcvi.org/ and http://genetics.bwh.harvard.edu/pph2/, respectively), were used to evaluate predictive functional consequences of nonsynonymous changes.
Three French families were included in an exome sequencing study with the following criteria: at least three individuals affected in a family, and at least one patient affected with multiple melanomas. Two affected relatives per family were sequenced. None of them carried CDKN2A, CDK4 or BAP1 mutations. Sequencing was done with the V4 Agilent kit exon capture (Otogenetics Company), coverage 50X, according to the manufacturer’s protocol using HiSeq 2000 from Illumina. Sequencing was carried out using 5 mcg of DNA per reaction. Variants were filtered using the DNA Nexus software (www.dnanexus.com). Only variants with allele frequency
Genes involved in the WNT and vesicular trafficking pathways are associated with melanoma predisposition Maider Ibarrola-Villava1, Rajiv Kumar2, Eduardo Nagore3, Meriem Benfodda4,5, Mickael Guedj6, Steven Gazal7, Hui-Han Hu4,5, Jian Guan2, P. Sivaramakishna Rachkonda2, Vincent Descamps4,8, Nicole Basset-Seguin4,8, Armand Bensussan4, Martine Bagot4,9, Philippe Saiag10, Dirk Schadendorf11, Manuel Martin-Gonzalez12, Matias Mayor13, Bernard Grandchamp5, Gloria Ribas1 and Nadem Soufir4,5 1
Department of Haematology and Medical Oncology, Biomedical Research Institute INCLIVA, Valencia 46010, Spain Division of Molecular Genetic Epidemiology, German Cancer Research Center, Heidelberg 69120, Germany 3 Department of Dermatology, Instituto Valenciano de Oncologia, Valencia 46009, Spain 4 Inserm U976, Centre de Recherche Sur la Peau, Hopital Saint Louis, Universite Paris 7, Paris 75010, France 5 ^pital Bichat, APHP, Paris 75018, France Departement de Genetique, Ho 6 Laboratoire Statistiques et Genomes, Evry 91000, France 7 UMR S738, Faculte de Medecine Xavier Bichat, Paris 75018, France 8 Department of Dermatology, Hopital Bichat, APHP, Paris 75018, France 9 Department of Dermatology, Hopital Saint Louis, APHP, Paris 75010, France 10 Department of Dermatology, Hopital Ambroise Pare, APHP, Paris 92100, France 11 Department of Dermatology, University Hospital Essen, Esse 45147, Germany 12 Department of Dermatology, Hospital Ramon y Cajal, Madrid 28034, Spain 13 Department of Dermatology, Hospital La Paz, Madrid 28046, Spain
Multifactorial predisposition to melanoma includes genes involved in pigmentation, immunity and DNA repair. Nonetheless, missing heritability in melanoma is still important. We studied the role of 335 candidate SNPs in melanoma susceptibility by using a dedicated chip and investigating 110 genes involved in different pathways. A discovery set was comprised of 1069 melanoma patients and 925 controls from France. Data were replicated using validation phases II (1085 cases and 801 controls from Spain) and III (1808 cases and 1894 controls from Germany and a second set of Spanish samples). In addition, an exome sequencing study was performed in three high-risk French melanoma families. Nineteen SNPs in 17 genes were initially associated with melanoma in the French population. Six SNPs were replicated in phase II, including two new SNPs in the WNT3 (rs199524) and VPS41 (rs11773094) genes. The role of VPS41 and WNT3 was confirmed in a meta-analysis (3940 melanoma cases and 3620 controls) with two-side p values of 0.002, (OR 5 0.86) and 4.07 3 10210 (OR 5 0.80), respectively. Exome sequencing revealed a non-synonymous VPS41 variant in one family that was shown to be strongly associated with familial melanoma (OR 5 4.46, p 5 0.001) in an independent sample of 178 melanoma families. WNT3 belongs to WNT pathway known to play a crucial role in melanoma, whereas VPS41 regulates vesicular trafficking and is thought to play a role in pigmentation. Our work identified two new pathways involved in melanoma predisposition. These results may be useful in the future for identifying individuals highly predisposed to melanoma.
The knowledge of the genetic predisposition to melanoma has significantly increased during the last several years. Linkage studies, candidate gene approaches, genome-wide association studies (GWAS) and whole-exome sequencing studies have
identified several loci that influence susceptibility to cutaneous malignant melanoma (MM).1–3 Mutations in the highly penetrant CDKN2A (MIM#60160), CDK4 (MIM#609048), TERT (MIM#187270) and POT1 (MIM#606478) genes are well
Key words: exome sequencing, case-control study, WNT3, VPS41, melanoma Abbreviations: CI: confidence interval; GWAS: genome-wide association studies; HWE: Hardy Weinberg equilibrium; MAF: minor allele frequency; MM: malignant melanoma; OR: odds ratio; SNP: single-nucleotide polymorphism Additional Supporting Information may be found in the online version of this article. Grant sponsor: Ministerio de Salud Carlos III; Grant number: PI10-0405, CP08-00069; Grant sponsor: La Societe Francaise de Dermatologie (SFD), FEDER DOI: 10.1002/ijc.29257 History: Received 7 Apr 2014; Accepted 9 Sep 2014; Online 10 Oct 2014 Correspondence to: Nadem Soufir, Departement de Genetique, H^opital Bichat Claude Bernard, 46 rue Henri Huchard, 75018 Paris, France. Tel.: [33140258551], Fax: 1[33140258785], E-mail: nadem.soufi[email protected]
C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
Cancer Genetics
2
2
VPS41 and WNT3 and melanoma susceptibility
Cancer Genetics
What’s new? While our understanding of the molecular pathways that lead to malignant melanoma has increased over the past few years, it is still incomplete. In this study, the authors identified two new pathways: one that involves WNT signaling (involved in melanoblast development), and another that involves vesicular trafficking (involved in changes in pigmentation phenotype). These results may be useful in developing screening tests for people who are genetically susceptible to melanoma. In addition, the whole-exome sequencing techniques used in this study may aid researchers in identifying mutations in unknown diseasecandidate genes.
known to play a role in familial melanoma, which accounts for about 10% of all MM cases.4–6 Moreover, low penetrance genetic variants have been associated with sporadic form of the disease and, therefore, pigmentation, nevus count, immune responses, DNA repair, oxidative stress and the vitamin D receptor have been associated with melanoma susceptibility.7–12 Although different genetic pathways are involved in melanoma onset, the human pigmentation pathway seems to play a critical role in the pathogenesis of the disease. MM predisposition is increased in individuals characterized by fair skin colour, high mole count, blond or red hair and high sun sensibility (Fitzpatrick’s phototypes I and II).4 Since the incidence of malignant melanoma in part reflects the consequences of a complex interaction between skin phenotype and UV exposure, variations in genes implicated in pigmentation and tanning response to UV are associated with susceptibility to MM. Thus, comprehensive knowledge of the entire network of genetic variants in pigmentation and melanogenesis genes may be a useful predictor of MM risk in the general population.13,14 The pigmentation pathway is a multistep process that involved differentiation and migration of melanocytes (pigment-producing cells), melanosome biogenesis, and migration and melanin synthesis. More than 120 genes have been identified through animal models. Nevertheless, only one gene, the melanocortin 1 receptor (MC1R, MIM#155555) is known to account for a substantial variation in the incidence of sporadic MM risk.15 Several other genes, such as OCA2 (MIM#611409), ASIP (MIM#600201), SLC45A2 (MIM#606202), TYR (MIM#606933) or TYRP1 (MIM#115501) have also emerged as being potentially important in melanoma susceptibility.1, 16–18 Considering the involvement of pigmentation in MM predisposition, the goal of the current study was to identify novel variants associated with MM risk by using singlenucleotide polymorphism (SNP) array on three independent sporadic melanoma series of European samples: France, Germany and Spain. Moreover, we performed whole-exome sequencing on a series of three melanoma families.
Material and Methods Study subjects and data collection Discovery phase I. A total number of 1069 melanoma cases
were recruited from 2002 to 2008 at the Departments of Dermatology of five different university-affiliated hospitals from
the Paris region (Saint Louis Hospital, Bichat Hospital, Tarnier Hospital, Ambroise Pare Hospital and Tenon Hospital). This cohort, called MelanCohort, comprised 888 sporadic melanoma cases (83%), 61 multiple melanoma patients (5.7%) and 120 nonrelated familial melanoma cases (11.3%). Similarly, 925 skin cancer-free control subjects were recruited from the EFS (Etablissement Franc¸ais du Sang) at Bichat and Saint Louis hospitals (Paris region) during the same period. A standardised questionnaire was used to collect information on pigmentation characteristics, the presence of childhood sunburns, Fitzpatrick’s skin type classification and personal and family history of cancer. Validation phase II. A second phase of the study was com-
posed of 1085 nonrelated MM cases obtained from the Departments of Dermatology of three different hospitals in Madrid, Spain (Gregorio Mara~ non Hospital, La Paz Hospital and Ramon y Cajal Hospital) and from the Instituto Valenciano de Oncologia in Valencia, Spain. A total of 801 cancerfree controls from the college of lawyers (Madrid, Spain), Gregorio Mara~ non Hospital and Instituto Valenciano de Oncologia were recruited. All subjects came from the geographical regions covered by the hospitals involved in the study. Validation phase III. A third phase of the study was com-
posed of two different sample sets. The first sample set included 872 nonrelated MM cases and 1158 cancer-free controls obtained from a different collection from the Instituto Valenciano de Oncologia. The second one included 936 MM patients and 736 controls recruited from the German Cancer Research Centre (Heidelberg, Germany). A total of 1808 melanoma cases and 1894 control subjects were included in this phase III validation. All subjects came from the geographical regions covered by their respective study hospitals. All enrolled patients and controls were from Caucasian origin. Distribution of cases and controls in discovery and validation phases is summarised in Supporting Information Table 1 and Supporting Information Table 2. Genomic DNA from cases and controls was isolated from peripheral blood lymphocytes using the MagNA Pure LC Instrument according to the manufacturer’s protocol (Roche Applied Science, Mannheim, Germany). DNA concentration was quantified using Quant-iT PicoGreen dsDNA Reagent (Invitrogen, Eugene, OR). All DNA samples were subjected to a quality control before processing them on the Illumina C 2014 UICC Int. J. Cancer: 00, 00–00 (2014) V
3
Ibarrola-Villava et al.
arrays (Illumina, San Diego, CA). All subjects gave written informed consent, and the study was approved by the Ethics Committees of the Gregorio Mara~ non Hospital, Hospital Clinico Universitario of Valencia and the Ambroise Pare Hospital (Boulogne, France).
quality control measure, one sample duplicate and a nontemplate sample per 96-well plate were included. For the first validation set (Phase II), a total of 13 SNPs were genotyped while 7 SNPs were studied in the second validation sample set (Phase III). Genotypes were scored by the LGC Genomics Company.
Gene and SNP selection
Exome sequencing
Previous literature and the information available on public databases were used for gene selection. Eventually, 110 genes were included in the study. The most overrepresented category with 96 genes was pigmentation. Pigmentation genes were chosen from the albinism and pigmentation genes databases (http://www.espcr.org/micemut/). A further group of 14 genes implicated in DNA repair (3 genes), cell cycle (5 genes), immune responses and apoptosis (4), MAP kinase pathway (1 gene) and glutamate receptor (1 gene) were included in the study. Three databases were used to collect information about single nucleotide polymorphisms (SNPs): NCBI (www.ncbi. nlm.nih.gov/SNP), HapMap (www.hapmap.org) and Illumina databases (www.illumina.com). Finally, 384 SNPs were chosen that included (a) polymorphisms predicted to have a functional impact (snpinfo.niehs.nih.gov/snpinfo/snpfunc.htm) that comprise SNPs located in putative transcription factor binding sites (66 SNPs) or in conserved regions (81 SNPs), nonsynonymous polymorphisms (49 SNPs), SNPs predicted to affect splicing (65 SNPs) or miRNA binding sites (63 SNPs), and (b) some tag-SNPs (87 SNPs). Tag-SNPs range from the hypothetical promoter area (2000 Kb upstream) until 2000 Kb downstream of the gene. Eight SNPs commonly associated with MM were included as positive controls (rs4911414 in ASIP, rs11547464, rs1805007 and rs1805009 in MC1R, rs16891982 in SLC45A2, rs1126809 in TYR, rs2733832 in TYRP1 and rs910873 in the 20Q region). To ensure a high genotyping success rate, a minor allele frequency threshold of 0.05 in the European HapMap population and a recommended manufacturer “Illumina score” were established in the SNP selection. SIFT and PolyPhen, (http:// sift.jcvi.org/ and http://genetics.bwh.harvard.edu/pph2/, respectively), were used to evaluate predictive functional consequences of nonsynonymous changes.
Three French families were included in an exome sequencing study with the following criteria: at least three individuals affected in a family, and at least one patient affected with multiple melanomas. Two affected relatives per family were sequenced. None of them carried CDKN2A, CDK4 or BAP1 mutations. Sequencing was done with the V4 Agilent kit exon capture (Otogenetics Company), coverage 50X, according to the manufacturer’s protocol using HiSeq 2000 from Illumina. Sequencing was carried out using 5 mcg of DNA per reaction. Variants were filtered using the DNA Nexus software (www.dnanexus.com). Only variants with allele frequency
