Cell Oncogenic Signaling Pathways in The Cancer Genome Atlas Article Graphical Abstract 10 Cancer Pathways Hippo Nrf2 PI3K/Akt Myc Notch EGFR 4% ERBB2 RTKs 5% r1 KRAS 9% v-' NF1 RAS 1 5% -< j RAF Mutation II: i f- 9,125 Tumor Samples 33 Cancer Types H633 H80 487 717 119 jXUjXE 4 981 Methylation XOK J-Ll i I U- Expression Copy Number Fusions EGFR 4% KRAS 9% BRAF 7% 'MAPK Pathway RTK RAS RAF Highlights • Alteration map of 10 signaling pathways across 9,125 samples from 33 cancer types • Reusable, curated pathway templates that include a catalogue of driver genes • 57% of tumors have at least one potentially actionable alteration in these pathways • Co-occurrence of actionable alterations suggests combination therapy opportunities Authors Francisco Sanchez-Vega, Marco Mina, Joshua Armenia.....Giovanni Ciriello, Chris Sander, Nikolaus Schultz Correspondence Giovanni.Ciriello@unil.ch (G.C.), sander.research@gmail.com (C.S.), schultz@cbio.mskcc.org (N.S.) In Brief An integrated analysis of genetic alterations in 10 signaling pathways in >9,000 tumors profiled by TCGA highlights significant representation of individual and co-occurring actionable alterations in these pathways, suggesting opportunities for targeted and combination therapies. Sanchez-Vega et al., 2018, Cell 173, 321-337 April 5, 2018 © 2018 Elsevier Inc. https://doi.Org/10.101 e/j.cell.2018.03.035 Cell Article Cell Oncogenic Signaling Pathways in The Cancer Genome Atlas Francisco Sanchez-Vega,1'230 Marco Mina,430 Joshua Armenia,12'30 Walid K. Chatila,1 Augustin Luna,5 Konnor C. La, Sofia Dimitriadoy,6 David L. Liu,7 Havish S. Kantheti,8 Sadegh Saghafinia,4 Debyani Chakravarty,1 Foysal Daian,1 Qingsong Gao,9 Matthew H. Bailey,9 Wen-Wei Liang,9 Steven M. Foltz,9 llya Shmulevich,10 Li Ding,911 Zachary Heins,1 Angelica Ochoa,1 Benjamin Gross,1 Jianjiong Gao,1 Hongxin Zhang,1 Ritika Kundra,1 Cyriac Kandoth,1 Istemi Bahceci,12 Leonard Dervishi,12 Ugur Dogrusoz,12 Wanding Zhou,13 Hui Shen,13 Peter W. Laird,13 Gregory P. Way,24 Casey S. Greene,24 Han Liang,25 Yonghong Xiao,26 Chen Wang,27 Antonio lavarone,28 Alice H. Berger,14 Trever G. Bivona,15 Alexander J. Lazar,16 Gary D. Hammer,17 Thomas Giordano,18 Lawrence N. Kwong,19 (Author list continued on next page) 1 Marie-Josee and Henry R. Kravis Center for Molecular Oncology, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA 2Human Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA departments of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA department of Computational Biology, University of Lausanne (UNIL), 1011 Lausanne, Vaud, Switzerland and Swiss Institute of Bioinformatics (SIB), Lausanne, Switzerland 5cBio Center, Dana-Farber Cancer Institute, Boston, MA; Department of Cell Biology, Harvard Medical School, Boston, MA 6Princeton University, Princeton, NJ, USA 7Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA; Cancer Program, Broad Institute of MIT and Harvard, Cambridge, MA, US 8University of Texas at Dallas, Richardson, TX 75080, USA department of Medicine and McDonnell Genome Institute, Washington University in St. Louis, St. Louis, Missouri, 63110, USA 10lnstitute for Systems Biology, Seattle, WA, USA "Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO 63110, USA 12Computer Engineering Department, Bilkent University, Ankara 06800, Turkey 13Van Andel Research Institute, 333 Bostwick Ave NE, Grand Rapids Michigan, 49503, USA 14Human Biology Division, Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA 15UCSF Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, 1450 3rd Street, San Francisco, California 94143, USA 1 departments of Pathology, Genomic Medicine & Translational Molecular Pathology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd-Unit 85, Houston, Texas 77030, USA (Affiliations continued on next page) SUMMARY Genetic alterations in signaling pathways that control cell-cycle progression, apoptosis, and cell growth are common hallmarks of cancer, but the extent, mechanisms, and co-occurrence of alterations in these pathways differ between individual tumors and tumor types. Using mutations, copy-number changes, mRNA expression, gene fusions and DNA methylation in 9,125 tumors profiled by The Cancer Genome Atlas (TCGA), we analyzed the mechanisms and patterns of somatic alterations in ten canonical pathways: cell cycle, Hippo, Myc, Notch, Nrf2, PI-3-Kinase/Akt, RTK-RAS, TGF(3 signaling, p53 and p-catenin/Wnt. We charted the detailed landscape of pathway alterations in 33 cancer types, stratified into 64 subtypes, and identified patterns of co-occurrence and mutual exclusivity. Eighty-nine percent of tumors had at least one driver alteration in these pathways, and 57% percent of tumors had at least one alteration potentially targetable by currently available drugs. Thirty percent of tumors had multiple targetable alterations, indicating opportunities for combination therapy. INTRODUCTION Cancer is a disease in which cells have acquired the ability to divide and grow uncontrollably (Hanahan and Weinberg, 2000, Hanahan and Weinberg, 2011), usually through genetic alterations in specific genes. Advances in DNA sequencing over the past decade have made it possible to systematically study these genetic changes, and we now have a better understanding of the commonly involved processes and signaling pathways (Garr-away and Lander, 2013; Vogelstein et al., 2013). As more genetic alterations become targetable by specific drugs, DNA sequencing is becoming part of routine clinical care (Hartmaier et al., 2017; Schram et al., 2017; Sholl et al., 2016; Zehir et al., 2017). However, there is considerable variation in the genes and pathways altered across different tumor types and individual tumor samples, and a complete understanding of the genes and Cell 773, 321-337, April 5, 2018 © 2018 Elsevier Inc. 321 This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/). Cell Grant McArthur,2029 Chenfei Huang,21 Aaron D. Tward,22 Mitchell J. Frederick,21 Frank McCormick,23 Matthew Meyerson,7 and The Cancer Genome Atlas Research Network, Eliezer M. Van Allen,7'31 Andrew D. Cherniack,7'31 Giovanni Ciriello,4-31 * Chris Sander,531* Nikolaus Schultz1233132* 17Department of Internal Medicine, Division of Metabolism, Endocrinology and Diabetes, Endocrine Oncology Program, University of Michigan, Ann Arbor, Michigan, Ml 48105, USA 18Department of Pathology, University of Michigan Medical School, Ann Arbor, Ml; Department of Internal Medicine, Division of Metabolism, Endocrinology & Diabetes, University of Michigan Medical School, Ann Arbor, Ml; Comprehensive Cancer Center, Michigan Medicine, Ann Arbor, Ml, USA 19Department of Translational Molecular Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA 20Peter MacCallum Cancer Centre, Melbourne, VIC, Australia 21 Dept. of Otolaryngology, Baylor College of Medicine, USA 22University of California, San Francisco Department of Otolaryngology-Head and Neck Surgery. 2233 Post Street, San Francisco, CA, 94143, USA 23UCSF Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, 1450 3rd Street, San Francisco, CA 94143, USA 24Department of Systems Pharmacology and Translational Therapeutics, University of Pennsylvania, Philadelphia, PA, 19104, USA 25Department of Bioinformatics and Computational Biology, The University of Texas MD Anderson Cancer Center, Houston, TX, 77030, USA 26TESARO Inc., Waltham, MA, 02451, USA 27Department of Health Sciences Research and Department of Obstetrics and Gynecology, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN, 55905, USA 28lnstitute for Cancer Genetics, Department of Neurology and Department of Pathology and Cell Biology, Columbia University Medical Center, New York, NY, 10032, USA 29University of Melbourne, Melbourne, VIC, Australia 30These authors contributed equally 31 Co-senior author 32Lead contact 'Correspondence: Giovanni.Ciriello@unil.ch (G.C.), sander.research@gmail.com (C.S.), schultz@cbio.mskcc.org (N.S.) https://doi.Org/10.1016/j.cell.2018.03.035 pathways altered in all cancer types is essential to identify potential therapeutic options and vulnerabilities. Several important signaling pathways have been identified as frequently genetically altered in cancer, including the RTK/ RAS/MAP-Kinase (hereafter also called RTK-RAS for brevity) pathway, PI3K/Akt signaling, and others (Vogelstein and Kinzler, 2004) . Members of these pathways and their interactions have been captured in a number of pathway databases, such as Pathway Commons (Cerami et al., 2011), which aggregates a number of databases, including REACTOME (Joshi-Tope et al., 2005) and KEGG (Kanehisa and Goto, 2000). Genes in key pathways are not altered at equal frequencies, with certain genes recurrently altered and well-known in cancer, while others are only rarely or never altered. The detection of recurrence of rare alterations often requires large numbers of samples (Lawrence et al., 2014). This is confounded by the challenge to distinguish between functionally relevant (or "driver" alterations) and non-oncogenic "passenger" events (Gao et al., 2014), especially in tumor types with a high background mutation burden (Alexandrovetal., 2013; Lawrence et al., 2013). In these cases, many mutations, even when they occur in cancer genes, may have no functional effect. This topic is further addressed in Bailey et al. (Bailey et al., 2018). Previous studies by The Cancer Genome Atlas (TCGA) have incrementally mapped out the alteration landscape in signaling pathways. Certain pathways, such as RTK-RAS signaling or the cell-cycle pathway, are altered at high frequencies across many different tumor types, whereas other pathways are altered in more specific subsets of malignancies (e.g., alterations in the oxidative stress response pathway are strongly associated with squamous histologies). With >10,000 samples characterized by TCGA, there is an opportunity to systematically characterize and define the alterations within well-known cancer pathways across all tumor types and map out commonalities and differences across pathways. The existence of shared genomic features across histologies has been highlighted by several studies (Ciriello et al., 2013; Hoadley et al., 2014,2018), but these studies traditionally used a gene-centric, as opposed to pathway-centric, approach. Identifying relationships of inter- and intra-pathway recurrence, co-occurrence or mutual exclusivity across different types of cancers can help elucidate functionally relevant mechanisms of oncogenic pathway alterations that might inform treatment options. Here, we worked within the framework of the TCGA PanCancer Atlas initiative (Cancer Genome Atlas Research Network et al., 2013c) to build a uniformly processed dataset and a unified data analysis pipeline aimed at exploring similarities and differences in canonical cancer pathway alterations across 33 cancer types. The focus of this effort is on mitogenic signaling pathways with vidence for functional alterations; other oncogenic processes, such as alterations in DNA repair (Knijnenburg et al., 2018), the spliceosome (Seiler et al., 2018), ubiquitination (Ge et al., 2018), or metabolic pathways (Peng et al., 2018), as well as the effects of splicing mutations (Jayasinghe et al., 2018), are covered by other efforts of the TCGA PanCancer Atlas project. RESULTS Dataset We evaluated all samples in the TCGA PanCancer Atlas collection for which the following data types were available: somatic mutations (whole-exome sequencing), gene expression levels 322 Cell 773, 321-337, April 5, 2018 Cell BRCA Normal (n=36) Her2-enriched (n=78) Basal (n=171) LumB (n=197) LumA (n=499) MESO (n=82) LUSC (n=464) LUAD (n=502) THCA (n=480) PCPG (n=161) ,ACC (n=76) CRC POLE (n=10) GS (n=58) MSI (n=63) CIN (n=328) STES POLE (n=9) EBV (n=30) GS (n=51) MSI (n=75) ESCC (n=90) CIN (n=297) CHOL (n=36) LIHC (n=348) PAAD (n=152) BLCA (n=399) KICH (n=65) HNSC HPV- (n=415) HPV+ (n=72) UVM (n=80) IDHmut-non-codel (n=248) IDHmut-codel (n=167) IDHwt (n=92) GBM (n=126) THYM (n=119) LAML (n=162) DLBC (n=37) LMS (n=83) MFS/UPS (n=80) DDLPS (n=46) Other (n=20) SKCM (n=363) LGG SARC KIRC (n=352) KIRP (n=271) PRAD (n=479) UCS (n=56) CN HIGH (n=163) CN LOW (n=147) -MSI (n=148) POLE (n=49) UCEC OV (n=177) Squamous Carcinoma (n=229)|QpgQ Adenocarcinoma (n=43) | Non-seminoma (n=82) I jqqj -Seminoma (n=62) | Uniformly processed genomic data for TCGA PanCancer Atlas Pathway databases Scientific literature & review articles Pathway analyses in prior TCGA publications -BS. ■•i ■ .■■r-^ Initial definition of pathway templates Definition of Known driver mutations OnC@KB driver alterations TSG / OG labels Recurrently altered genes (MutSig, Lawrence et at, TCGA Copy-Number Portal) Known functional gene fusions/rearrangements Preliminary t * pathway templates Recurrently mutated positions Cancer Hot spots I 3D Hotspots Epigenetic silencing calls (RESET + manual curation) Removal of non-altered genes Cu rated pathway templates Pathway-specific expert or analysis working group review Pathway curation Relevant somatic alterations in tabular format Data analysis and visualization tools: cBioPortal SELECT Ü PathwayMapper Final pathway templates and analysis Figure 1. TCGA PanCancer Atlas Pathways Dataset and Workflow (A) Distribution of cancer types in the cohort, including molecular subtypes analyzed. (B) Workflow for pathway curation and analysis. Genes were curated from previous TCGA efforts and the scientific literature. Only genes with evidence for statistically recurrent or known driver alterations in the uniformly processed TCGA PanCancer Atlas dataset were included in the curated pathway templates. TCGA disease codes and abbreviations: AML, acute myeloid leukemia; ACC, adrenocortical carcinoma; BRCA, breast cancer; CESC, cervical cancer; KICH, chromophobe renal cell carcinoma; KIRC: clear cell kidney carcinoma; CRC, colorectal adenocarcinoma; SKCM, cutaneous melanoma; DLBC, diffuse large B cell lymphoma; GBM, glioblastoma multiforme; HNSC, head and neck squamous cell carcinoma; LIHC, liver hepatocellular carcinoma; LGG, lower grade glioma; (legend continued on next page) Cell 773, 321-337, April 5, 2018 323 Cell (RNA-Seq), DNA copy-number alterations (Affymetrix SNP6 arrays), and DNA methylation (Infinium arrays). This resulted in a final set of 9,125 samples from 33 different cancer types (Figure 1A, Table S1). In order to account for molecular or histological subtypes, these cancer types were further stratified into a total of 64 genomically distinct tumor subtypes, as previously defined by the individual TCGA analysis working groups (Figure 1A, Table S1). All genomic data and clinical attributes per sample can be visualized through the cBioPortal for Cancer Genomics at http://www.cbioportal.org/ (Cerami etal., 2012). Definition of Pathways and Alterations We evaluated 10 canonical signaling pathways with frequent genetic alterations, starting with key cancer genes explored in these pathways in previous TCGA publications, and focused on pathway members likely to be cancer drivers (functional contributors) or therapeutic targets. The pathways analyzed are: (1) cell cycle, (2) Hippo signaling, (3) Myc signaling, (4) Notch signaling, (5) oxidative stress response/Nrf2, (6) PI-3-Kinase signaling, (7) receptor-tyrosine kinase (RTK)/RAS/MAP-Kinase signaling, (8) TGFF3 signaling, (9) p53 and (10) F3-catenin/Wnt signaling (Figures 2 and S1, Table S2). Alterations in DNA repair pathways, epigenetic modifiers, splicing, and other cellular processes frequently altered in cancer were not included, as these primarily provide a background of genomic instability, rather than specifically proliferative potential. We began by compiling and reviewing the full set of cancer-type specific pathway diagrams from the compendium of TCGA manuscripts published between 2008 and 2017 (Brennan et al., 2013; Cancer Genome Atlas Network, 2012a, 2012b, 2015a, 2015b, Cancer Genome Atlas Research Network, 2008, 2011, 2013a, 2013b, 2014a, 2014b, 2014c, 2014d, 2017a, 2017b; Davis et al., 2014), each of which included the pathway genes found to be genetically altered in the individual tumor types. These pathway diagrams are publicly available as predefined network templates within the www.PathwayMapper. org visualization tool (Bahceci et al., 2017). By taking the union of pathway members across multiple TCGA studies, we produced a consolidated list of candidate member genes for each of the ten pathways. These were then further curated based on updated literature (including but not limited to the references in Table S2), public pathway databases, and expert opinion (Figure 1B). The selected genes in the ten pathways were then assessed for recurrent alterations within and across different tumor types as follows (Figure 1B): Alterations of pathway members were classified as activating events (usually specific recurrent missense mutations, i.e., hotspot mutations, amplifications, or fusions involving oncogenes) or inactivating events (truncating mutations, specific recurrent missense or inframe mutations, deletions, as well as fusions and promoter hypermethylation of tumor suppressor genes). Individual alterations were also scrutinized for two features: statistical recurrence across sets of tumor samples and presumed functional impact. We first assessed statistical recurrence using MutSigCV (Lawrence et al., 2014) for mutations and GISTIC 2.0 (Mermel et al., 2011) for copy-number alterations. In order to identify likely functional variants, we then used recurrence across tumor samples at the residue level (linear and 3D mutational hotspots; Chang et al., 2016, 2018; Gao et al., 2017; see STAR Methods) and prior knowledge about specific variants via the OncoKB knowledge base, which contains information about the oncogenic effects and treatment implications of variants in > 400 cancer genes (Chakravarty et al., 2017a). Epigenetic silencing through promoter DNA hypermethylation of tumor suppressor genes was evaluated using the RESET algorithm (see STAR Methods). Gene fusions and structural rearrangements were called from RNA-Seq data using a combination of the STAR-Fusion, EricScript and BreakFast algorithms (Gao et al., 2018, see STAR Methods), and likely passenger events were filtered out based on OncoKB annotation. Through this process, genes without evidence for recurrent or previously known oncogenic alterations were removed from the preliminary pathway templates. The resulting curated pathway templates and the identified genetic alterations were vetted for functional importance by individual pathway experts or the corresponding TCGA PanCancer Atlas pathway-specific analysis working groups, when applicable (Figure 1B). The pathway member genes and the genetic alterations considered as oncogenic are listed in Table S3, and binary genomic alteration matrices are provided as Table S4 (see STAR Methods). The resulting comprehensive dataset of different types of alterations across many tumor types form the basis of all subsequent analyses regarding pathways, patterns of co-occurrence and mutual exclusivity, as well as potential therapeutic implications. The simplified pathway diagrams in Figure 2 show the most frequently altered genes in the ten pathways, including alteration frequencies as well as the types of oncogenic alterations identified in each of the genes. Pathway Alteration Frequencies per Tumor Type For each tumor type and subtype, we computed the fraction of samples with at least one alteration in each of the 10 signaling pathways (Figure 3). A tumor sample was considered as altered in a given pathway if one or more genes in the pathway contained a recurrent or known driver alteration (as described above). Despite the fact that non-recurrent and not previously known alterations were filtered out as likely passenger events and were not included in the alteration frequencies, the microsatellite instability (MSI) and polymerase e (POLE) mutant subtypes of LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; OV, ovarian serous cystadenocarcinoma; KIRP, papillary kidney carcinoma; THCA, papillary thyroid carcinoma; ST AD, stomach adenocarcinoma; PRAD, prostate adenocarcinoma; BLCA, urothelial bladder cancer; UCS, uterine carcinosarcoma; UCEC, uterine corpus endometrial carcinoma; ESCA, esophageal cancer; PCPG, pheochromocytoma & paraganglioma; PAAD, pancreatic ductal adenocarcinoma; MESO, mesothelioma; UVM, uveal melanoma; SARC, sarcoma; CHOL, cholangiocarcinoma; TGCT, testicular germ cell cancer; THYM, thymoma; STES, stomach and esophageal cancer; EBV, Epstein-ESarrvirus; HPV, human papillomavirus; DDLPS, dedifferentiated liposarcoma; LMS, leiomyosarcoma; MFS/UPS, myxofibrosarcoma/undifferentiated pleomorphic sarcoma; ESCC, esophageal squamous cell carcinoma; GS, genomically stable; CIN, chromosomal instability; MSI, microsatellite instability. 324 Cell 773, 321-337, April 5, 2018 Cell RTK/RAS pathway I EGFR ] \ ERBB2 ERBbJ] I ERBB4 11 MET | [ PDGFRA I FGFR1 ]| FGFR2 |[~FGFR3 | [~FGFR4 ]| KIT 11 IGF1R | I RET II ROS1 |(~ ALK || FLT3 11 NTRK1-3|| JAK2~| X_, I RTKs NF1 J |_RASA11 —»| MAPK1-] <— J MAP2Kl"| | MAP2K2| MEK Proliferation Cell survival Translation Wnt pathway I SFRP1-5|—I I WNTligands ] RNF43 |i ZNRF3 |- I WNT Dual Receptor Complex -1 Hippo pathway ~r—r i I TEAD2 I I Cell proliferation and differentiation Nrf2 pathway KEAP1 II CUL3 NFE2L2 —► Oxidative Stress TGFß pathway TGFß ligands Activin ligands TGFBR1/2| I ACVR2A/1B -'-\-'- SMAD2 II SMAD3 | 1 Proliferation, stem/progenitor phenotype Myc pathway 1^ J I ^MYCN JI MYCL MLX/MONDO complex PI3K pathway 1 I IPIK3CA/B I- I PIK3R1/3 [->| STK11 |-'—I RICTOR J —► J TSC1/2 |—I [ŘP J J mTORCI Cell growth MDM2/4 I DNA replication stress Oncogenic 1 I Cell survival, proliferation Cell cycle pathway [ CDKN1A/B ICDKN2A/B/C I CCNE1 J —' 9 Cyclin/CDKs E2F1/3 i Cell cycle Notch pathway JAG2 |[aRRDC1| "T" 1 NOV NOTCH1 NOTCH2 NOTCH3 NOTCH4 Cleaved NOTCH EP300 -c HES-X Cell growth, —»-Activation —I Inhibition I I Part of complex Copy number changes Mutations Fusions/Rearrangements-Epigenetic silencing' Gene Alteration frequencies Oncogene □1%Q>5%B>10% Tumor suppressor n<1% Q >1%Q >5%| >10% Figure 2. Curated Pathways Pathway members and interactions in the ten selected pathways. Genes are altered at different frequencies (color intensity indicates the average frequency of alteration within the entire dataset) by oncogenic activations (red) and tumor suppressor inactivations (blue). The types of somatic alteration considered for each gene (copy-number alterations, mutations, fusions or epigenetic silencing) are specified using a set of four vertical dots on the left of each gene symbol. An expanded version including cross-pathways interactions is provided as Figure S1. gastrointestinal and uterine tumors, which had the highest mutation burden, also had the highest overall frequencies of pathway alterations. This is possibly due to the frequent inactivating mutations introduced by the predominant mutation mechanisms in these tumor types (Boland and Goel, 2010; Rayner etal., 2016). The RTK-RAS pathway was the signaling pathway with the highest median frequency of alterations (46% of samples) Cell 773, 321-337, April 5, 2018 325 Cell Alteration frequencies GBM 77 86 57 48 18 8 6 10 2 119 0.37 4.3 LGG IDHwt 82 64 47 29 27 5 1 5 68 0.25 4.3 LGG IDHmut-codel 9 45 22 5 26 66 3 99 50 34 0.21 0.8 LGG IDHmut 19 28 15 92 24 8 10 92 21 1 60 0.17 0.8 UVM 6 6 4 2 10 1 2 10 1 65 0.28 0.4 HNSC HPV+ 26 32 60 11 25 8 4 8 11 1 83 0.32 3.3 HNSC HPV- 45 86 39 82 36 16 20 42 13 13 108 0.43 4.6 84 14 4 1 4 13 2 1 2 16 0.03 0.4 22 30 16 28 11 41 7 5 1 146 0.78 1.8 32 15 6 6 11 10 1 4 1 81 0.33 0.3 THYM 14 9 4 7 5 7 1 7 3 2 26 0.09 0.6 LUAD 56 38 61 21 19 23 23 10 15 118 0.48 8.2 MESO 9 54 13 21 9 6 7 40 2 98 0.41 0.8 LUSC 54 79 68 86 31 18 12 28 11 25 158 0.61 7.7 BRCALumA 28 31 62 25 14 15 12 5 4 1 101 0.34 2.0 BRCALumB 44 48 48 49 25 31 26 15 10 2 211 0.60 2.0 BRCA Her2-enriched 82 40 60 78 18 17 29 10 8 1 230 0.53 4.3 BRCA Basal 46 51 53 91 38 11 39 14 8 4 246 0.67 2.7 BRCA Normal 36 36 33 31 3 6 19 3 53 0.16 1.5 STES Squamous 50 89 53 96 38 13 22 21 13 23 189 0.59 3.1 STES CIN 63 74 33 76 21 26 21 16 23 2 222 0.58 3.4 STES EBV 50 100 80 13 83 67 7 10 17 52 0.22 4.1 STES GS 31 39 18 24 31 20 12 4 20 2 66 0.10 2.1 STES MSI-POLE 71 64 64 49 79 70 19 54 57 2 85 0.19 37.1 CRC MSI-POLE 99 74 68 49 74 95 52 64 55 1 45 0.09 56.9 CRC GS 88 45 53 19 29 90 21 10 38 5 55 0.23 2.9 CRC CIN 66 36 32 84 23 91 17 8 22 1 115 0.54 2.9 LIHC 22 69 25 37 26 43 19 12 7 7 121 0.45 2.9 CHOL 56 53 17 19 8 17 19 17 3 6 100 0.58 1.8 PAAD 78 70 19 69 14 12 14 7 41 62 0.26 3.5 KIRC 14 14 17 6 8 7 5 5 3 3 49 0.25 1.6 KIRP 17 12 8 4 12 9 6 11 1 6 49 0.35 2.2 KICH 5 23 15 32 3 3 2 3 5 77 0.80 0.9 BLCA 64 46 62 42 20 18 26 9 9 150 0.50 6.8 PRAD 15 28 32 21 13 35 11 5 6 1 92 0.16 1.3 TGCT sem 63 8 11 6 6 2 70 0.54 0.4 TGCT non-sem 20 7 5 5 16 2 10 2 99 0.67 0.4 OV 58 48 49 96 28 10 40 21 5 5 316 0.79 2.4 UCEC CN high 61 43 86 90 32 18 31 13 5 5 296 0.67 1.9 UCEC CN low 37 9 95 10 14 54 10 7 1 5 42 0.15 2.1 UCEC MSI-POLE 71 31 98 42 64 70 30 55 31 19 30 0.08 71.2 UCS 61 70 79 91 54 18 27 16 4 4 247 0.71 3.5 CESC Adeno 63 21 56 19 30 14 16 14 21 5 95 0.36 3.6 CESC Squamous 32 19 59 12 35 12 5 33 11 10 101 0.44 5.2 SKCM 94 77 33 28 27 23 10 25 7 1 131 0.53 22.1 SARC DDLPS 43 83 20 85 17 15 7 9 7 450 0.36 1.1 SARC LMS 31 55 33 71 14 11 4 4 1 4 177 0.69 1.8 SARC MFS/UPS 48 74 32 68 34 20 8 21 6 4 328 0.66 3.0 SARC other 25 30 15 5 5 5 10 104 0.33 1.2 DLBC 24 76 8 19 70 70 14 35 14 90 0.29 3.5 LAML 49 17 3 9 18 11 2 3 1 1 28 0.05 1.1 46 45 33 29 23 15 11 10 7 1 $ t 'S" # 4 / > > O >- 0.24; Mina et al., 2017) included. (H) Breakdown of the interactions involving EGFR amplifications and mutations, corresponding to the bounding boxes in panel G. Left side: mutually exclusive interactions. Right side: co-occurring interactions. Cell 773, 321-337, April 5, 2018 331 .CD Cell number of alterations in a subset of tumor types are currently biomarkers for standard care targeted therapies, and a larger number are potential biomarkers for investigational therapies, some with promising clinical results. Using the OncoKB knowledge base of clinically actionable alterations (Chakravarty et al., 2017a), we systematically assessed all alterations in each sample of each cancer type, distinguishing between standard care actionability (Levels 1 or 2) and investigational therapies (Levels 3 and 4). Overall, 51 % of tumors had at least one potentially actionable alteration in the ten signaling pathways, and 57% had at least one actionable alteration when including genes outside of these pathways, most notably BRCA1/2 and IDH1/2 (all numbers referenced below include these additional genes). Apart from the Her2-enriched breast cancer samples, most of which have a standard care targeted therapy, melanoma was the tumor type with the highest fraction of tumors with a Level 1 or 2A alteration (46%) (Figure 7A), mainly due to frequent BRAF mutations (Figure 7B), followed by esophagogastric cancers {ERBB2 amplifications). Luminal A breast cancer was the tumor type with the highest frequency of biomarkers with promising investigational data (Level 3A), driven by the high prevalence of PIK3CA, AKT1 and ERBB2 mutations. Several tumor types had frequent mutations that are biomarkers for drug sensitivity in other cancer types (Level 3B), including endometrial cancer, where PIK3CA mutations are common. Uveal melanoma and testicular non-seminoma had the lowest percentage of potentially targetable samples (2.5% and 8.5%, respectively); thymoma, mesothelioma (MESO), and renal clear cell carcinoma (KIRC) also had low frequencies of potentially actionable alterations. Thirty percent of tumor samples had two or more potentially targetable alterations (Figure 7C). Among these, the MSI-H and POLE-mutated tumor subtypes had the highest proportion of samples with multiple potentially actionable alterations (not considering the fact that patients with MSI-H tumors are now eligible for immunotherapy). Other tumor types with a high frequency of samples with multiple targetable alterations included non-hypermutated endometrial cancer (64%), colorectal cancer (37%), and breast cancer (28%). Finally, we searched for candidate drug combinations that could prove effective across different tumor types based on the occurrence of actionable alterations. Hypermutant MSI and POLE subtypes had a high fraction of samples of actionable alterations corresponding to various drug combinations. In other tumor subtypes, a combination of CDK4 and MDM2 inhibitors was the most commonly indicated combination (1 % total), in particular in dedifferentiated liposarcomas (SARC DDLPS), in which 78% of the cases had co-amplification of the two targets (Figure 7D). By a similar consideration linking actionable alterations of targets to their inhibitors, a combination of HER2 and PI3K inhibitors might be beneficial across multiple tumor types, in particular Her2-enriched breast cancer (17%), uterine carcinosarcoma (UCS, 7%), chromosomally unstable endometrioid carcinoma (UCEC CN high, 7%), and cervical adenocarcinoma (7%) (Figure 7D). Additional candidate combination therapies include PI3K and MEK inhibitors in EBV+ stomach tumors (10%), CDK4 and PI3K inhibitors in glioblastoma multiforme (7%), HER2 and MEK inhibitors in pancreatic cancer (7%), PI3K and RAF inhibitors in melanoma (SKCM, 12%), and IDH and PI3K inhibitors in IDH-mutant low grade glioma (14%) (Figure 7D). While there are many steps from the observation of combinations of genetic alterations to valid combination therapies, this survey indicates the wide landscape of potential tumor-type specific novel therapeutic combinations that can be explored in experimental and clinical contexts. DISCUSSION Signaling pathways are somatically altered in cancer at varying frequencies and in varying combinations across different organs and tissues, indicative of complex interplay and pathway crosstalk. Understanding the extent, detailed mechanisms, and co-occurrence of the oncogenic alterations in these pathways is critical for the development of new therapeutic approaches that can improve patient care. Here we performed a comprehensive characterization of 10 selected signaling pathways across the 33 cancer types analyzed by TCGA. This report constitutes the first pan-cancer exploration that uses a uniformly processed dataset and a standardized set of pathway templates, curated through a combination of computational methods and expert review (Figures 1 and 2). The results highlight similarities and differences in frequencies of alteration of individual pathways in different cancer types and specific molecular subtypes (Figure 3). They also underscore the potential for discovering previously uncharacterized alterations in pathway genes that occur at low frequencies and might otherwise remain statistically unnoticeable (see SOS1, Figure 4). More generally, even though a small set of critical genes contains a very large fraction of alterations in these pathways (Figures 4 and 5), there is a complex interplay of co-occurring and mutually exclusive alterations within and across pathways (Figure 6). In spite of the accumulating wealth of biological knowledge and the accepted oncogenic relevance of these pathways, the number of currently approved biomarkers Figure 7. Therapeutic Actionability and Drug Combinations (A) Frequencies of clinical actionability by cancer subtype, broken down by level of evidence (Levels 1-4). Samples are classified by the alteration that carries the highest level of evidence. Tumor type-specific samples are analyzed by variants considered actionable, oncogenic but not actionable, or variants of unknown significance (VUS). (B) Frequencies of actionable alterations per gene across cancer subtypes. For genes with different levels for different alterations, multiple rows are shown. Genes are grouped by pathway. Six additional genes not in the ten pathways (BRCA1, BRCA2, ERCC2, IDH1, IDH2, ESR1) are included and taken into account in the overall frequencies. (C) Fraction of samples with a given number of actionable alterations per tumor type. (D) Frequencies of possible drug combinations indicated by the co-alteration of actionable variants in each tumor type for the most frequent drug class combinations. Cell 773, 321-337, April 5, 2018 333 Cell linked to standard of care therapies remains sparse (Figure 7), but additional drug targets in these pathways will hopefully emerge, and candidates for combination therapy will be explored. This analysis of targetable alterations only included currently approved therapies or investigational therapies with reported promising results. These predominantly target the RTK-RAS, PI3K, cell-cycle, and p53 pathways. While some of these therapies are standard care, many are still investigational, and further testing is required to assess how effective different targeted therapies will be across tumor types and in tumors with different co-mutation spectra. Efforts are underway to develop therapies that target additional pathways, some of which are in clinical trials (Table S6) (Park and Guan, 2013), (Whitfield et al., 2017), (Whitfield et al., 2017), (Aster and Blacklow, 2012), (Takebe et al., 2014), (Buijs et al., 2012), (Sheen et al., 2013), (Pai et al., 2017). In the Wnt signaling pathway, for example, two approaches involve drugs targeting PORCN, which is involved in the processing of wingless proteins, and monoclonal antibodies directed at proteins in the Frizzled gene family. While the Nrf2 pathway does not have therapies directly targeting any of the pathway members included in this study, alterations in Nrf2 pathway members {NFE2L2 and KEAP1) are used as part of the inclusion criteria in the Phase 2 trial of a TORC1/2 inhibitor. Clinical trials involving these pathways exemplify opportunities in precision medicine to associate additional functional alterations as part of inclusion criteria (Table S6). Not all apparently functional mutations, however, represent therapeutic targets, as illustrated, e.g., by the unusually large number of mutations in the MSI-H and POLE-mutated tumor subtypes, of which only a small fraction plausibly dominate oncogenesis. The observed co-occurrence patterns indicate a potential for combination therapies in some tumor types. The development of targeted combination therapies has been challenging for several reasons, including lack of safety data for combinations, the relatively slow pace of adoption of clinically approved multi-panel gene tests and of clinical trials testing combinations of multiple targeted therapies. However, there is a growing corpus of promising preclinical data indicating such combinations can be effective, such as the combination of MDM2 and CDK4 inhibitors (Laroche-Clary et al., 2017), and the combination of PI3K inhibitors and HER2 inhibitors in HER2-positive/P/K3C/A mutant breast cancer patients, even when single gene-therapy approaches (e.g., PI3K monotherapy for PIK3CA mutant tumors) have thus far not had definitive clinical impact. Although we cover a diverse range of oncogenic processes that spans most tissues and organ systems (Figures 1 and 3), we did not include some tumor types in the scope of this TCGA project, including most hematologic cancers. Furthermore, in spite of the relatively large set of samples, this effort is still underpowered to reliably discover tumor-type specific alterations that occur at very low frequencies; these will require further exploration using larger tumor-type specific sample sets. The original aim and scope of TCGA was to genomically characterize primary, untreated tumors with a basic set of genetic alterations and transcript profiles. As the program is now completed, a future challenge is to expand these analyses to larger sample sets, additional data types, such as metabo- lite levels, a wider range of epigenetic states, post-transla-tional modifications of proteins, and to investigate metastatic disease and genomic alterations that arise in post-treatment samples, as well as analyzing the role of a wider range of germline alterations and their interplay with somatic events. These new avenues of research will benefit from pathway-level analysis for which the templates and template curation pipelines presented here constitute a promising starting point. Similarly, as the catalog of clinically actionable alterations continues to grow, understanding intra- and inter-pathway dependencies, such as the ones considered here, will be crucial for the development of effective combination therapies that address or prevent resistance to initially successful single agent therapies. The curated pathway templates and the uniformly processed dataset of alteration calls in 9,125 tumor samples are publicly available (Tables S3 and S4) and can be easily accessed through the PathwayMapper tool (http://pathwaymapper.org/), which allows alteration frequencies to be visually overlaid on the pathway templates; and, via the cBioPortal for Cancer Genomics (http://www.cbioportal.org/). This pathway landscape in The Cancer Genome Atlas is meant to provide a valuable resource for clinical oncologists, for cancer researchers and for a broad scientific community interested in cancer precision medicine. STAR* METHODS Detailed methods are provided in the online version of this paper and include the following: • KEY RESOURCES TABLE • CONTACT FOR REAGENT AND RESOURCE SHARING • EXPERIMENTAL MODEL AND SUBJECT DETAILS O Sample Selection and Exclusions • METHOD DETAILS o Somatic mutation calling o Pathway Template Curation o Epigenetic silencing o Gene fusion detection and filtering o Generation of Genomic Alteration Matrices (GAMs) o Analysis of conditional selection between alterations o Pathway-level analysis of conditional selection o Curation of Clinical Trials • QUANTIFICATION AND STATISTICAL ANALYSIS • DATA AND SOFTWARE AVAILABILITY • ADDITIONAL RESOURCES SUPPLEMENTAL INFORMATION Supplemental Information includes four figures and eight tables and can be found with this article online at https://doi.Org/10.1016/j.cell.2018.03.035. ACKNOWLEDGMENTS This work was supported by NIH Grants U54 HG003273, U54 HG003067, U54 HG003079, U24 CA143799, U24 CA143835, U24 CA143840, U24 CA143843, U24 CA143845, U24 CA143848, U24 CA143858, U24 CA143866, U24 CA143867, U24 CA143882, U24 CA143883, U24 CA144025, P30 CA016672. 334 Cell 773, 321-337, April 5, 2018 Cell AUTHORS CONTRIBUTIONS F.S.-V., M.M., J.A., W.K.C., A.L., E.M.V.A., A.D. C, G.C., C. S. and N. S. designed the study; F.S.-V., M.M., J.A., W.K.C., A.L., J.G., K.L., S.D., H.S.K., Z.H., A.O., B.G., J.G., H.Z., R.K., I.B., L.D., U.D., C.K., Q.G., M.H.B,W.-W.L, S.M.F., I.S., H.L, L.D., A.D.C., G.C., and N.S. collected and annotated the data; F.S.-V., M.M., J.A., W.K.C., A.L., D.C., W.Z., H.S., P.W.L., F.D., D.L.L., S.S., G.P.W., C.S.G., Y.X., C.W., A.I., A.H.B., T.G.B., A.J.L., G.D.H., T.G., L.N.K., G.M., CH., A.D.T., M.J.F., F.M., M.M., E.V.A, A.D.C., G.C., CS. and N.S. analyzed and interpreted the data. The Cancer Genome Atlas Research Network was responsible for the project administration. F.S.-V., M.M., J.A., W.K.C, A.L., E.M.V.A., A.D.C, G.C, CS. and N.S. drafted the manuscript. All authors participated in editing or reviewing of the manuscript, and all authors approved the submitted manuscript. DECLARATION OF INTERESTS Michael Seiler, Peter G. Smith, Ping Zhu, Silvia Buonamici, and Lihua Yu are employees of H3 Biomedicine, Inc. Parts of this work are the subject of a patent application: WO2017040526 titled "Splice variants associated with neomorphic sf3b1 mutants." Shouyoung Peng, Anant A. Agrawal, James Palacino, and Teng Teng are employees of H3 Biomedicine, Inc. Andrew D. Cherniack, Ashton C. Berger, and Galen F. Gao receive research support from Bayer Pharmaceuticals. Gordon B. Mills serves on the External Scientific Review Board of Astrazeneca. Anil Sood is on the Scientific Advisory Board for Kiyatec and is a shareholder in BioPath. Jonathan S. Serody receives funding from Merck, Inc. Kyle R. Covington is an employee of Castle Biosciences, Inc. Preethi H. Gunaratne is founder, CSO, and shareholder of NextmiRNA Therapeutics. Christina Yau is a part-time employee/consultant at NantOmics. Franz X. Schaub is an employee and shareholder of SEngine Precision Medicine, Inc. Carla Grandori is an employee, founder, and shareholder of SEngine Precision Medicine, Inc. Robert N. Eisenman is a member of the Scientific Advisory Boards and shareholder of Shenogen Pharma and Kronos Bio. Daniel J. Wei-senberger is a consultant for Zymo Research Corporation. Joshua M. Stuart is the founder of Five3 Genomics and shareholder of NantOmics. Marc T. Goodman receives research support from Merck, Inc. Andrew J. Gentles is a consultant for Cibermed. Charles M. Perou is an equity stock holder, consultant, and Board of Directors member of BioClassifier and GeneCentric Diagnostics and is also listed as an inventor on patent applications on the Breast PAM50 and Lung Cancer Subtyping assays. Matthew Meyerson receives research support from Bayer Pharmaceuticals; is an equity holder in, consultant for, and Scientific Advisory Board chair for OrigiMed; and is an inventor of a patent for EGFR mutation diagnosis in lung cancer, licensed to LabCorp. Eduard Porta-Pardo is an inventor of a patent for domainXplorer. Han Liang is a shareholder and scientific advisor of Precision Scientific and Eagle Nebula. Da Yang is an inventor on a pending patent application describing the use of antisense oligonucleotides against specific IncRNA sequence as diagnostic and therapeutic tools. Yonghong Xiao was an employee and shareholder of TESARO, Inc. Bin Feng is an employee and shareholder of TESARO, Inc. Carter Van Waes received research funding for the study of IAP inhibitor ASTX660 through a Cooperative Agreement between NIDCD, NIH, and Astex Pharmaceuticals. Raunaq Malhotra is an employee and shareholder of Seven Bridges, Inc. Peter W. Laird serves on the Scientific Advisory Board for AnchorDx. Joel Tepper is a consultant at EMD Serono. Kenneth Wang serves on the Advisory Board for Boston Scientific, Microtech, and Olympus. Andrea Califano is a founder, shareholder, and advisory board member of Darwin-Health, Inc. and a shareholder and advisory board member of Tempus, Inc. Toni K. Choueiri serves as needed on advisory boards for Bristol-Myers Squibb, Merck, and Roche. Lawrence Kwong receives research support from Array BioPharma. Sharon E. Plön is a member of the Scientific Advisory Board for Baylor Genetics Laboratory. Beth Y. Karlan serves on the Advisory Board of Invitae. Received: November 17, 2017 Revised: February 28, 2018 Accepted: March 15, 2018 Published: April 5, 2018 REFERENCES Alaei-Mahabadi, B., Bhadury, J., Karlsson, J.W., Nilsson, J.A., and Larsson, E. (2016). Global analysis of somatic structural genomic alterations and their impact on gene expression in diverse human cancers. Proc. Natl. Acad. Sei. USA 7 73, 13768-13773. Alexandrov, L.B., Nik-Zainal, S., Wedge, D.C, Aparicio, S.A.J.R., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., Borresen-Dale, A.-L, et al.; Australian Pancreatic Cancer Genome Initiative; ICGC Breast Cancer Consortium; ICGC MMML-Seq Consortium; ICGC PedBrain (2013). Signatures of mutational processes in human cancer. Nature 500, 415-421. Aster, J.C, and Blacklow, S.C (2012). Targeting the Notch pathway: twists and turns on the road to rational therapeutics. J. Clin. Oncol. 30, 2418-2420. Bahceci, I., Dogrusoz, U., La, K.C, Babur, Ö., Gao, J., and Schultz, N. (2017). PathwayMapper: a collaborative visual web editor for cancer pathways and genomic data. Bioinformatics 33, 2238-2240. Bailey, M.H., Tokheim, C, Porta-Pardo, E., Sengupta, S., Betrand, D., Weer-asinghe, A., Colaprico, A., Kim, J., Wendl, M.C, Reardon, B., et al. (2018). Comprehensive Discovery and Characterization of Driver Genes and Mutations in Human Cancers. Cell 773, this issue, 371-385. Benelli, M., Pescucci, C, Marseglia, G., Severgnini, M., Torricelli, F., and Magi, A. (2012). Discovering chimeric transcripts in paired-end RNA-seq data by using EricScript. Bioinformatics 28, 3232-3239. Boland, CR., and Goel, A. (2010). Microsatellite instability in colorectal cancer. Gastroenterology 738, 2073-2087.e3. Brennan, C.W., Verhaak, R.G.W., McKenna, A., Campos, B., Noushmehr, H., Salama, S.R., Zheng, S., Chakravarty, D., Sanborn, J.Z., Berman, S.H., et al.; TCGA Research Network (2013). The somatic genomic landscape of glioblastoma. Cell 755, 462-477. Buijs, J.T., Stayrook, K.R., and Guise, T.A. (2012). The role of TGF-ß in bone metastasis: novel therapeutic perspectives. Bonekey Rep. 7, 96. Campbell, J.D., Alexandrov, A., Kim, J., Wala, J., Berger, A.H., Pedamallu, CS., Shukla, S.A., Guo, G., Brooks, A.N., Murray, B.A., et al.; Cancer Genome Atlas Research Network (2016). Distinct patterns of somatic genome alterations in lung adenocarcinomas and squamous cell carcinomas. Nat. Genet. 48, 607-616. Cancer Genome Atlas Network (2012a). Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330-337. Cancer Genome Atlas Network (2012b). Comprehensive molecular portraits of human breast tumours. Nature 490, 61-70. Cancer Genome Atlas Network (2015a). Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature 57 7, 576-582. Cancer Genome Atlas Network (2015b). Genomic Classification of Cutaneous Melanoma. Cell 767, 1681-1696. Cancer Genome Atlas Research Network (2008). Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455, 1061-1068. Cancer Genome Atlas Research Network (2011). Integrated genomic analyses of ovarian carcinoma. Nature 474, 609-615. Cancer Genome Atlas Research Network (2012). Comprehensive genomic characterization of squamous cell lung cancers. Nature 489, 519-525. Cancer Genome Atlas Research Network (2013a). Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 499, 43-49. Cancer Genome Atlas Research Network, Kandoth, C, Schultz, N., Cherniack, A.D., Akbani, R., Liu, Y., Shen, H., Robertson, A.G., Pashtan, I., Shen, R., Benz, C.C., etal. (2013b). Integrated genomic characterization of endometrial carcinoma. Nature 497, 67-73. Cancer Genome Atlas Research Network,, Weinstein, J.N., Collisson, E.A., Mills, G.B., Shaw, K.R., Ozenberger, B.A., Ellrott, K., Shmulevich, I., Sander, C, and Stuart, J.M. (2013c). The Cancer Genome Atlas Pan-Cancer analysis project. Nat. Genet. 45, 1113-1120. Cell 773, 321-337, April 5, 2018 335 Cell Cancer Genome Atlas Research Network (2014a). Comprehensive molecular characterization of urothelial bladder carcinoma. Nature 507, 315-322. Cancer Genome Atlas Research Network (2014b). Comprehensive molecular profiling of lung adenocarcinoma. Nature 57 7, 543-550. Cancer Genome Atlas Research Network (2014c). Comprehensive molecular characterization of gastric adenocarcinoma. Nature 513, 202-209. Cancer Genome Atlas Research Network (2014d). Integrated genomic characterization of papillary thyroid carcinoma. Cell 159, 676-690. Cancer Genome Atlas Research Network (2017a). Comprehensive and Integrative Genomic Characterization of Hepatocellular Carcinoma. Cell 169, 1327-1341 .e23. Cancer Genome Atlas Research Network; Analysis Working Group: Asan University; BC Cancer Agency; Brigham and Women's Hospital; Broad Institute; Brown University; Case Western Reserve University; Dana-Farber Cancer Institute; Duke University; Greater Poland Cancer Centre; Harvard Medical School; et-al (2017b). Integrated genomic characterization of oesophageal carcinoma. Nature 541, 169-175. Cerami, E., Gao, J., Dogrusoz, U., Gross, B.E., Sumer, S.O., Aksoy, B.A., Ja-cobsen, A., Byrne, C.J., Heuer, M.L., Larsson, E., et al. (2012). The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2, 401-404. Cerami, E.G., Gross, B.E., Demir, E., Rodchenkov, I., Babur, O., Anwar, N., Schultz, N., Bader, G.D., and Sander, C. (2011). Pathway Commons, a web resource for biological pathway data. Nucleic Acids Res. 39, D685-D690. Chakravarty, D., Gao, J., Phillips, S.M., Kundra, R., Zhang, H., Wang, J., Rudolph, J.E., Yaeger, R., Soumerai, T., Nissan, M.H., et al. (2017a). OncoKB: A Precision Oncology Knowledge Base. JCO Precis. Oncol. 2017. Chakravarty, D., Pedraza, A.M., Cotari, J., Liu, A.H., Punko, D., Kokroo, A., Huse, J.T., Altan-Bonnet, G., and Brennan, C.W. (2017b). EGFR and PDGFRA co-expression and heterodimerization in glioblastoma tumor sphere lines. Sei. Rep. 7, 9043. Chang, M.T., Asthana, S., Gao, S.P., Lee, B.H., Chapman, J.S., Kandoth, C, Gao, J., Socci, N.D., Solit, D.B., Olshen, A.B., et al. (2016). Identifying recurrent mutations in cancer reveals widespread lineage diversity and mutational specificity. Nat. Biotechnol. 34, 155-163. Chang, M.T., Bhattarai, T.S., Schram, A.M., Bielski, CM., Donoghue, M.T.A., Jonsson, P., Chakravarty, D., Phillips, S., Kandoth, C, Penson, A., et al. (2018). Accelerating Discovery of Functional Mutant Alleles in Cancer. Cancer Discov. S, 174-183. Ciriello, G., Miller, M.L, Aksoy, B.A., Senbabaoglu, Y., Schultz, N., and Sander, C. (2013). Emerging landscape of oncogenic signatures across human cancers. Nat. Genet. 45, 1127-1133. Davis, C.F., Ricketts, C.J., Wang, M., Yang, L, Cherniack, A.D., Shen, H., Bu-hay, C, Kang, H., Kim, S.C., Fahey, C.C., et al.; The Cancer Genome Atlas Research Network (2014). The somatic genomic landscape of chromophobe renal cell carcinoma. Cancer Cell 26, 319-330. Ellrott, K., Bailey, M.H., Saksena, G., Covington, K.R., Kandoth, C, Stewart, C, Hess, J., Ma, S., McLellan, M., Sofia, HJ., et al. (2018). Scalable Open Science Approach for Mutation Calling of Tumor Exomes Using Multiple Genomic Pipelines. Cell Syst 6. https://doi.Org/10.1016/j.cels.2018.03.002. Etemadmoghadam, D., Weir, B.A., Au-Yeung, G., Alsop, K., Mitchell, G., George, J., Davis, S., D'Andrea, A.D., Simpson, K., Hahn, W.C., and Bowtell, D. D.; Australian Ovarian Cancer Study Group (2013). Synthetic lethality between CCNE1 amplification and loss of BRCA1. Proc. Natl. Acad. Sei. USA 7 70, 19489-19494. FANTOM Consortium and the RIKEN PMI and CLST (DGT), Forrest, A.R., Ka-waji, H., Rehli, M., Baillie, J.K., de Hoon, M.J., Haberle, V., Lassmann, T., Ku-lakovskiy, I.V., Lizio, M., Itoh, M., et al. (2014). A promoter-level mammalian expression atlas. Nature 507, 462-470. Gao, J., Ciriello, G., Sander, C, and Schultz, N. (2014). Collection, integration and analysis of cancer genomic profiles: from data to insight. Curr. Opin. Genet. Dev. 24, 92-98. Gao, J., Chang, M.T., Johnsen, H.C., Gao, S.P., Sylvester, B.E., Sumer, S.O., Zhang, H., Solit, D.B., Taylor, B.S., Schultz, N., and Sander, C. (2017). 3D clusters of somatic mutations in cancer reveal numerous rare mutations as functional targets. Genome Med. 9, 4. Gao, Q., Liang, W.-W., Foltz, S.M., Mutharasu, G., Jayasinghe, R.G., Cao, S., Liao, W.-W., Reynolds, S.M., Wyczalkowski, M.A., Yao, L, et al. (2018). Driver Fusions and Their Implications in the Development and Treatment of Human Cancers. Cell Rep 23. https://doi.Org/10.1016/j.celrep.2018.03.050. Garraway, L.A., and Lander, E.S. (2013). Lessons from the cancer genome. Cell 153, 17-37. Ge, Z., Leighton, J., Wang, Y., Peng, X., Chen, Z., Chen, H., Sun, Y., Yao, F., Li, J., Zhang, H., et al. (2018). Integrated Genomic Analysis of the Ubiquitin Pathway across CancerTypes. Cell Rep 23. https://doi.Org/10.1016/j.celrep. 2018.03.047. Hanahan, D., and Weinberg, R.A. (2000). The hallmarks of cancer. Cel 100, 57-70. Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next generation. Cell 144, 646-674. Hartmaier, R.J., Albacker, L.A., Chmielecki, J., Bailey, M., He, J., Goldberg, M.E., Ramkissoon, S., Suh, J., Elvin, J.A., Chiacchia, S., et al. (2017). High-Throughput Genomic Profiling of Adult Solid Tumors Reveals Novel Insights into Cancer Pathogenesis. Cancer Res. 77, 2464-2475. Hoadley, K.A., Yau, C, Wolf, D.M., Cherniack, A.D., Tamborero, D., Ng, S., Leiserson, M.D.M., Niu, B., McLellan, M.D., Uzunangelov, V., et al.; Cancer Genome Atlas Research Network (2014). Multiplatform analysis of 12 cancer types reveals molecular classification within and across tissues of origin. Cell 758, 929-944. Hoadley, K.A., Yau, C, Hinoue, T., Wolf, M.D., Lazar, J.A., Drill, E., Shen, R., Taylor, M.A., Cherniack, D.A., Akbani, R., et al. (2018). Cell-of-Origin Patterns Dominate the Molecular Classification of 10,000 Tumors from 33 Types of Cancer. Cell 773. https://doi.Org/10.1016/j.cell.2018.03.022. Jayasinghe, R.G., Cao, S., Gao, Q., Wendl, M.C., Vo, N.S., Reynolds, S.M., Zhao, Y., Climente-Gonzälez, H., Chai, S., Wang, F., et al. (2018). Systematic analysis of splice site-creating mutations in cancer. Cell Rep 23. https://doi. org/10.1016/j.celrep.2018.03.052. Joshi-Tope, G., Gillespie, M., Vastrik, I., D'Eustachio, P., Schmidt, E., de Bono, B., Jassal, B., Gopinath, G.R., Wu, G.R., Matthews, L, et al. (2005). Reactome: a knowledgebase of biological pathways. Nucleic Acids Res. 33, D428-D432. Kanehisa, M., and Goto, S. (2000). KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 28, 27-30. Korkut, A., Zaidi, S., Kanchi, R., Berger, A., Robertson, G., Kwong, L., Datto, M., Roszik, J., Ling, S., Ravikumar, V., et al. (2017). A PanCancer molecular analysis of disruptions in the TGF-ß pathway. Cancer Cell. Laroche-Clary, A., Chaire, V., Algeo, M.-P., Derieppe, M.-A., Loarer, F.L., and Italiano, A. (2017). Combined targeting of MDM2 and CDK4 is synergistic in dedifferentiated liposarcomas. J. Hematol. Oncol. 70, 123. Lawrence, M.S., Stojanov, P., Polak, P., Kryukov, G.V., Cibulskis, K., Siva-chenko, A., Carter, S.L., Stewart, C, Mermel, C.H., Roberts, S.A., et al. (2013). Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 499, 214-218. Lawrence, M.S., Stojanov, P., Mermel, C.H., Robinson, J.T., Garraway, L.A., Golub, T.R., Meyerson, M., Gabriel, S.B., Lander, E.S., and Getz, G. (2014). Discovery and saturation analysis of cancer genes across 21 tumour types. Nature 505, 495-501. Lepri, F., De Luca, A., Stella, L., Rossi, C, Baldassarre, G., Pantaleoni, F., Cor-deddu, V., Williams, B.J., Dentici, M.L, Caputo, V., et al. (2011). SOS1 mutations in Noonan syndrome: molecular spectrum, structural insights on pathogenic effects, and genotype-phenotype correlations. Hum. Mutat. 32, 760-772. Mermel, C.H., Schumacher, S.E., Hill, B., Meyerson, M.L, Beroukhim, R., and Getz, G. (2011). GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol. 72, R41. 336 Cell 773, 321-337, April 5, 2018 Cell Mina, M., Raynaud, F., Tavernari, D., Battistello, E., Sungalee, S., Saghafinia, S., Laessle, T., Sanchez-Vega, F., Schultz, N., Oricchio, E., and Ciriello, G. (2017) . Conditional Selection of Genomic Alterations Dictates Cancer Evolution and Oncogenic Dependencies. Cancer Cell 32, 155-168.e6. Mitsuishi, Y., Taguchi, K., Kawatani, Y., Shibata, T., Nukiwa, T., Aburatani, H., Yamamoto, M., and Motohashi, H. (2012). Nrf2 redirects glucose and gluta-mine into anabolic pathways in metabolic reprogramming. Cancer Cell 22, 66-79. Nissan, M.H., Pratilas, C.A., Jones, A.M., Ramirez, R., Won, H., Liu, C, Tiwari, S., Kong, L, Hanrahan, A.J., Yao, Z., et al. (2014). Loss of NF1 in cutaneous melanoma is associated with RAS activation and MEK dependence. Cancer Res. 74, 2340-2350. Pai, S.G., Carneiro, B.A., Mota, J.M., Costa, R., Leite, CA., Barroso-Sousa, R., Kaplan, J.B., Chae, Y.K., and Giles, F.J. (2017). Wnt/beta-catenin pathway: modulating anticancer immune response. J. Hematol. Oncol. 10, 101. Park, H.W., and Guan, K.-L. (2013). Regulation of the Hippo pathway and mplications for anticancer drug development. Trends Pharmacol. Sei. 34, 581-589. Peng, X., Chen, Z., Farshidfar, F., Xu, X., Lorenzi, P., Wang, Y., Cheng, F., Tan, L., Mojumdar, K., Du, D., et al. (2018). Molecular Characterization and Clinical Relevance of Metabolic Expression Subtypes in Human Cancers. Cell Rep 23. https://doi.Org/10.1016/j.celrep.2018.03.077. Rayner, E., van Gool, I.C., Palles, C, Kearsey, S.E., Bosse, T., Tomlinson, I., and Church, D.N. (2016). A panoply of errors: polymerase proofreading domain mutations in cancer. Nat. Rev. Cancer 16, 71-81. Schaub, F.X., Dhankani, V., Berger, A.C., Trivedi, M., Richardson, A.B., Shaw, R., Zhao, W., Zhang, X., Ventura, A., Liu, Y., et al. (2018). Alterations of the MYC oncogene and its Proximal Network across The Cancer Genome Atlas. Cell Syst 6. https://doi.Org/10.1016/j.cels.2018.03.003. Schräm, A.M., Berger, M.F., and Hyman, D.M. (2017). Precision oncology: Charting a path forward to broader deployment of genomic profiling. PLoS Med. 14, e1002242. Seiler, M., Peng, S., Agrawal, A.A., Palacino, J., Teng, T., Zhu, P., Smith, P.G., Cancer Genome Atlas Research Network, The, Buonamici, S., Yu, L., et al. (2018) . Survey of Splicing Factor Genes and Their Functional Consequences across 33 Cancer Types. Cell Rep 23. https://doi.Org/10.1016/j.celrep.2018. 01.088. Sheen, Y.Y., Kim, M.-J., Park, S.-A., Park, S.-Y., and Nam, J.-S. (2013). Target-ng the Transforming Growth Factor-ß Signaling in Cancer Therapy. Biomol. Ther. (Seoul) 21, 323-331. Sholl, L.M., Do, K., Shivdasani, P., Cerami, E., Dubuc, A.M., Kuo, F.C., Garcia, E.P., Jia, Y., Davineni, P., Abo, R.P., et al. (2016). Institutional implementation of clinical tumor profiling on an unselected cancer population. JCI Insight 1, e87062. Sun, C, Wang, L, Huang, S., Heynen, G.J.J.E., Prahallad, A., Robert, C, Haa-nen, J., Blank, C, Wesseling, J., Willems, S.M., et al. (2014). Reversible and adaptive resistance to BRAF(V600E) inhibition in melanoma. Nature 508, 118-122. Takebe, N., Nguyen, D., and Yang, S.X. (2014). Targeting notch signaling pathway in cancer: clinical development advances and challenges. Pharmacol. Ther. 141, 140-149. Knijnenburg, T., Wang, L., Zimmermann, M., Chambwe, N., Gao, G., Cher-niack, A., Fan, H., Shen, H., Way, G., Greene, C, et al. (2018). Genomic and Molecular Landscape of DNA Damage Repair Deficiency Across The Cancer Genome Atlas. Cell Rep 23. https://doi.Org/10.1016/j.celrep.2018.03.076. Triche, T.J., Jr., Weisenberger, D.J., Van Den Berg, D., Laird, P.W., and Sieg-mund, K.D. (2013). Low-level processing of lllumina Infinium DNA Methylation BeadArrays. Nucleic Acids Res. 41, e90. Unni, A.M., Lockwood, W.W., Zejnullahu, K., Lee-Lin, S.-Q., and Varmus, H. (2015). Evidence that synthetic lethality underlies the mutual exclusivity of oncogenic KRAS and EGFR mutations in lung adenocarcinoma. eLife 4, e06907. Vogelstein, B., and Kinzler, K.W. (2004). Cancer genes and the pathways they control. Nat. Med. 10, 789-799. Vogelstein, B., Papadopoulos, N., Velculescu, V.E., Zhou, S., Diaz, LA., Jr., and Kinzler, K.W. (2013). Cancer genome landscapes. Science 339, 1546-1558. Wang, Z., Yang, B., Zhang, M., Guo, W., Wu, Z., Wang, Y., Jia, L, Li, S., The Cancer Genome Atlas Research Network, Xie, W., and Yang, D. (2018). LncRNA epigenetic landscape analysis identifies EPIC1 as an oncogenic IncRNAthat interacts with MYC and promotes cell cycle progression in cancer. Cancer Cell 33. https://doi.Org/10.1016/j.ccell.2018.03.006. Way, G.P., Sanchez-Vega, F., La, K., Armenia, J., Chatila, W.K., Luna, A., Sander, C, Cherniack, A.D., Mina, M., Ciriello, G., et al. (2018). Pan-cancer Ras Pathway Activation in The Cancer Genome Atlas. Cell Rep 23. https:// doi.org/10.1016/j.celrep.2018.03.046. Whitfield, J.R., Beaulieu, M.-E., and Soucek, L. (2017). Strategies to Inhibit Myc and Their Clinical Applicability. Front. Cell Dev. Biol. 5, 10. Zehir, A., Benayed, R., Shah, R.H., Syed, A., Middha, S., Kim, H.R., Srinivasan, P., Gao, J., Chakravarty, D., Devlin, S.M., et al. (2017). Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat. Med. 23, 703-713. Zhang, Y., Kwok-Shing Ng, P., Kucherlapati, M., Chen, F., Liu, Y., Tsang, Y.H., de Velasco, G., Jeong, K.J., Akbani, R., Hadjipanayis, A., et al. (2017). A Pan-Cancer Proteogenomic Atlas of PI3K/AKT/mTOR Pathway Alterations. Cancer Cell37,820-832.e3. Cell 173, 321-337, April 5, 2018 337 Cell STAR* METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Deposited Data Raw and processed clinical, array and sequence data. NCI Genomic Data Commons https://portal.gdc.cancer.gov/ https://gdc.cancer.gov/about-data/publications/pancanatlas cBioPortal: http://www.cbioportal.org Digital Pathology Images Cancer Digital Slide Archive http://cancer.digitalslidearchive.net/ Software and Algorithms PathwayMapper (Bahceci et al., 2017) http://www.pathwaymapper.org/ SELECT (Mina et al., 2017) http://ciriellolab.org/select/select.html GISTIC 2.0 (Mermel et al., 2011) http://archive.broadinstitute.org/cancer/cga/gistic MutSigCV (Lawrence et al., 2014) http://software.broadinstitute.org/cancer/software/ genepattern/modules/docs/MutSigCV STAR-Fusion Hass et al., bioRxiv https://doi.Org/10.1101/120295 https://github.com/STAR-Fusion/STAR-Fusion/wiki Breakfast See link https://github.com/annalam/breakfast EricScript (Benellietal.,2012) https://sites.google.com/site/bioericscript/ RESET (Saghafinia, Mina et al., manuscript under review) http://ciriellolab.org/ Other OncoKB (Chakravarty et al., 2017a) www.oncokb.org Can cerHots pots (Chang et al., 2016) www.cancerhotspots.org 3D Hotspots (Gaoetal.,2017) www.3dhotspots.org cBioPortal Cerami et al., 2012 http://www.cbioportal.org/ TCGA Batch Effects The University of Texas MD Anderson Cancer Center http://bioinformatics.mdanderson.org/tcgambatch/ Pathway Commons (Cerami etal., 2011) http://www. path waycom mons .org/ CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Nikolaus Schultz (schultz@cbio.mskcc.org). EXPERIMENTAL MODEL AND SUBJECT DETAILS TCGA Project Management has collected necessary human subjects documentation to ensure the project complies with 45-CFR-46 (the "Common Rule"). The program has obtained documentation from every contributing clinical site to verify that IRB approval has been obtained to participate in TCGA. Such documented approval may include one or more of the following: • An IRB-approved protocol with Informed Consent specific to TCGA or a substantially similar program. In the latter case, if the protocol was not TCGA-specif ic, the clinical site PI provided a further finding from the IRB that the already-approved protocol is sufficient to participate in TCGA. • A TCGA-specif ic IRB waiver has been granted. • A TCGA-specific letter that the IRB considers one of the exemptions in 45-CFR-46 applicable. The two most common exemptions cited were that the research falls under 46.102(f)(2) or 46.101(b)(4). Both exempt requirements for informed consent, because the received data and material do not contain directly identifiable private information. • A TCGA-specific letter that the IRB does not consider the use of these data and materials to be human subjects research. This was most common for collections in which the donors were deceased. e1 Cell 773, 321-337.e1-e4, April 5, 2018 Cell Sample Selection and Exclusions We started from the set of 11,276 patients that were included in the final whitelist for the TCGA PanCanAtlas project. We only used samples that had available data across these four genomic platforms: mutations, copy number, DNAmethylation and mRNA expression. Our analyses excluded certain molecular platforms that have previously been used in TCGA, such as protein levels from reverse-phase protein arrays (RPPA), microRNA, and IncRNA, as their inclusion would have implied a sharp decrease in the total number of samples with data available across all platforms. Additionally, we excluded samples that had been flagged during pathology review by an expert committee or due to quality control (QC) issues identified by the individual tumor-type or PanCanAtlas analysis working groups. After these filters had been applied, a total of 9,125 patients were used. Samples consisted of primary solid tumors for a large majority of these patients (8602/9125,94%), plus a small number of blood tumors corresponding to the AML subset (162/9125, 2%) and a small subset of metastatic samples from melanoma patients (361/9125, 4%). METHOD DETAILS Somatic mutation calling We used version 2.8 of the mutation annotation format (MAF) file provided by the MC3 ("Multi-Center Mutation Calling in Multiple Cancers") group within the TCGA Network (Ellrott et al., 2018). The mutation data can be found here (https://gdc.cancer.gov/ about-data/publications/mc3-2017). We augmented this file in collaboration with the MC3 group and included all validated mutation calls from the original AML publication. The final MAF that was used for our analyses, including OncoKB annotations, can be dow-loaded from the TCGA PanCancer Atlas publication page (https://gdc.cancer.gov/about-data/publications/pancanatlas). Pathway Template Curation We manually curated the gene annotation of the ten selected pathways using the following workflow. Selection and classification of genes in pathways Genes were assigned to pathways based on a combined revision of pathway analyses in previous TCGA marker papers published between 2008 and 2017, a review of the scientific literature (including but not limited to the references in Table S2) and expert curation. We applied two different kinds of expert curation. 1) several of the pathways, such asTGF-Beta, Mycand PI3K, had specific analysis working groups. These groups were led by experts in each pathway and published separate manuscripts (Ge et al., 2018; Korkutetal., 2017; Pengetal., 2018; Schaubetal., 2018; Wang etal., 2018; Way and Greene, 2017). 2) for some of the pathways, we consulted experts from outside of TCGA in order to validate or improve our curated pathway templates (e.g., Frank McCormick for RTK-RAS or Mitchell Frederick for Notch). After the lists of pathway members were finalized, each gene was annotated as Tumor Suppressor (TSG) or Oncogene (OG) using OncoKB and prior knowledge from the scientific literature. The final gene lists that were selected for each pathway are provided in Table S3. Identification of mutational hotspots The cancer hotspots algorithm that we used identifies recurrent alterations based on a cohort of 24,592 tumor samples (Chang et al., 2016, 2018). Identification of 3D hotspots was based on recurrence of mutations in the context of spatial neighborhoods in protein structures (Gao et al., 2017). Annotation of functionally relevant mutations We used information about oncogenic and clinically actionable mutations from the OncoKB database (Chakravarty et al., 2017a), which provides information on variants in more than 400 genes. For template curation, we used OncoKB to filter out putative passenger mutations and copy number changes, by discarding somatic alterations that were not labeled as oncogenic, likely oncogenic or predicted oncogenic in the database. For the analysis of therapeutic implications, we used annotations about different levels of clinical actionability as described in the text. These had originally been compiled and curated by OncoKB by combining a diverse set of sources, including FDA-, NCCN- and other guidelines, ClinicalTrials.gov and the scientific literature. Annotation of functionally relevant CNVs We applied a two step procedure to determine whether the annotated genes were functionally amplified or deleted in each specific sample. First, a collection of functional relevant amplifications and deletions was curated by integrating the GISTIC 2.0 analysis of the PancanAtlas dataset and the OncoKB database. GISTIC was run separately on each individual tumor type, and then globally on the entire PanCanAtlas dataset, yielding a list of recurrently amplified and deleted regions of interest (ROIs). Default parameters of GISTIC 2.0 were used, with the confidence level set to 0.95. For Genes within ROIs, copy number variants consistent with the role of the gene (amplification of OGs and deletions of TSGs) were retained. Gene-specific copy number variants that were labeled as oncogenic, likely oncogenic or predicted oncogenic in OncoKB were also retained, yielding a list of gene-level functional CNVs. As an additional validation step, we individually inspected each of the gene level calls to ensure that there was a good correlation between copy-number status and gene expression, and we excluded calls in genes for which this correlation was non-existent. Thresholded gene-level amplification/deletion values produced by GISTIC were used for pathway analysis, considering only amplifications (+2) and deep deletions (-2). In total, 7,532 gene amplifications and 5,602 deletions were selected, for a total of 13,134 occurrences. Cell 773, 321-337.e1-e4, April 5, 2018 e2 Cell Epigenetic silencing Curated analysis of CDKN2A promoter hypermethylation CDKN2A promoter methylation was assessed using lllumina Infinium HumanMethylation450 probe cg13601799 located within Exon 1 a of CDKN2A (p16INK4a). We described the selection of this probe for CDKN2A methylation calling in a prior report (Cancer Genome Atlas Research Network, 2012). We introduced a further refinement of DNA methylation calling to avoid artifactual hypermethylation calls due to deep deletion in a gene. In brief, we used Level 1 IDAT files to calculate out-of-band ('oob') probe intensities as a surrogate for background intensity, superior to internal negative controls (Triche et al., 2013). cg13601799 is a Type I probe with both methylated (M) and unmethylated (U) versions in the red color channel, and therefore dye bias is not a concern. We compared the foreground intensities (M and U) to the empirical distribution of the background intensities (as measured by the 'oob' probes). We first called a sample to be methylated when the methylated (M) signal was higher than the 95th percentile of the background ('oob') probes (FDR = 5%). As this locus is unmethylated in normal tissues, the U signals are generally higher than the M signal due to the presence of contaminating normal cell types. We required a Log2(foreground/background) log-ratio of 2 or greater for the U probe to ensure that the U signal was derived from tumor cells and not from contaminating normal cells in the case of a tumor with CDKN2A deletion. If Log2(foreground/background) was < 2 for U and < Ofor M for this probe, then we concluded that this locus was deleted in the tumor cells, and we then denoted these cases as containing "no signal" (Table S4). We identified 681 such samples, and we had GISTIC copy number change data for 627 out of these 681. Out of these 627, 471 were called to have high-level deletion for CDKN2A (-2 in GISTIC calls) and 120 had low level deletion for this gene (-1 in GISTIC calls), validating this approach. Analysis of DNA hypermethylation at the promoters of other tumor suppressor genes Epigenetic DNA hypermethylation events at promoters of tumor suppressor genes that are associated with decreased gene expression were systematically identified using the RESET bioinformatic tool (Saghafinia, Mina et al. manuscript in preparation). RESET extracts probes that (i) map to gene promoter regions, (ii) are significantly hypermethylated compared to normal tissue samples, and (iii) are associated with lower transcript levels of the corresponding gene. More specifically, only probes overlapping gene promoter regions extracted from the FANTOM5 cohort of robust promoters are considered (FANTOM Consortium and the RIKEN PMI and CLST (DGT) et al., 2014). The status of a probe (dichotomized in hypermethylated and not hypermethylated) is determined by comparing its beta value to the beta value distribution from adjacent normal tissue samples available in the TCGA sample collection. Finally, RESET determines whether a hypermethylation event is associated with mRNAdownregulation by checking whether the mRNA expression of the associated gene is significantly decreased in hypermethylated tumors, compared to the not hypermethylated ones. To avoid biases due to intrinsic gene expression and methylation differences between tumor samples of different origins, we separately applied RESET within each tumor type. For tumor types without normal adjacent samples, the entire set of normal samples from the TCGA cohort was used to define the background beta value distribution. In this study, we evaluated all tumor suppressors in ten pathway templates (Table S3). We considered as significant only silencing events with a false discovery rate FDR < 10% and a RESET score > 1. The results were further manually curated to exclude cases where the methylation event might be tissue-associated, leading to a list of 15 genes silenced by DNA methylation (Table S7) Consistently with the procedure used for copy number calls, all hypermethylation occurrences for these 15 genes in all tumor samples were retained, even if the silencing event was only significantly recurrent in a subset of tumor types. This pancan set of occurrences was further filtered to increase the likelihood of functional relevance: only the hypermethylation occurrences with a gene expression lower than the 25 percentile of the gene expression distribution from the unmethylated samples were retained as functional and considered in the downstream analyses. The sample-specific epigenetic silencing calls are provided as part of the genomic alteration matrix described below (Table S4). Gene fusion detection and filtering TCGA RNA-Seq data were downloaded from Cancer Genomics Hub (CGHub, http://cghub.ucsc.edu) and analyzed using Google cloud. For each sample, the fastq file was mapped to the human genome (build 38) followed by fusion calling using STAR-Fusion (parameters:-annotation -coding-effect), EricScript (default parameters) and BREAKFAST (two different minimum distance cutoffs were used: 5 kb and 100 kb). STAR-Fusion showed higher sensitivity in detecting the fusions reported in previous TCGA studies. Therefore, we focused on the STAR-Fusion output and integrated EricScript and BREAKFAST output in one of the following filtering steps: 1) an exclusion list of genes was curated, including uncharacterized genes, immunoglobin genes, mitochondrial genes, etc. Fusions involving these genes were filtered; 2) Fusions from the same gene or paralog genes (downloaded from https://github.com/ STAR-Fusion/STAR-Fusion_benchmarking_data/tree/master/resources) were filtered; 3) Fusions reported in normal samples were filtered, including the ones from TCGA normal samples, GTEx tissues (reported in STAR-Fusion output), and non-cancer cell study (Babiceanu et al.); 4) For the fusions reported by only STAR-Fusion but not EricScript, a minimum value of FFPM (fusion fragments per million total reads) was required, as suggested by the author; For the fusions reported by both callers, no requirement. 5) Finally, fusions with exactly the same breakpoints in > 10 samples across different cancer types were removed unless they were reported in previous TCGA studies (e.g., FGFR3-TACC3). For our pathway analyses, we included only the fusions that (a) involved at least one gene labeled as TSG in one of our pathway templates, or (b) involved at least one gene labeled as OG in one of our pathway templates and such that the fusion is labeled as oncogenic, likely oncogenic or predicted oncogenic in OncoKB. We also included a small set of additional fusions (MAML3-UBTF, NOTCH2-SEC22B and PIK3CA-TBL1XR) based on recent evidence from the literature. Any fusion failing to satisfy at least e3 Cell 773, 321-337.e1-e4, April 5, 2018 Cell one of these requirements was excluded from subsequent pathway analyses (although some additional fusions that are clinically actionable based on OncoKB where included in Figure 7 for completeness). The final set of all fusion calls used in our manuscript is provided as Table S8. Generation of Genomic Alteration Matrices (GAMs) To integrate all the genomic data in a format readily usable in the downstream analyses, the complexity of mutation and CNV data was summarized into a binary Genomic Alteration Matrix (GAM) representing the occurrence of gene alterations across samples, provided as Table S4. This matrix includes the set of functionally relevant mutations and CNVs selected for each gene and summarized in the onco-query language column provided as part of each pathway template in Table S3. In the alteration level version of this matrix, copy number events and point mutation events affecting the same gene were kept distinct. We also included epigenetic silencing of CDKN2A based on DNA methylation analysis of the gene promoter and the epigenetic silencing of 15 additional genes uncovered by RESET. The resulting table has entries for 9,125 samples and 411 alterations, for a total of 33,324 occurrences. For completeness, in Table S4 we also provide a version of the GAM where alterations are aggregated at the gene level and a third version were alterations are aggregated at the pathway level for the ten pathways in our analysis. Analysis of conditional selection between alterations SELECT, a method that infers conditional selection dependencies between alterations from occurrence patterns (Mina et al., 2017), was run on the PancanPathway GAM. The default parameters of the R package implementation were used, with 5,000 random permutations. SELECT analysis was performed at alteration level, considering as separate features the point mutations, copy number changes, silencing and fusion events affecting the same gene. Alteration type, tumor type and tumor subtype were used as covariates in the analysis. Only alterations with more than 5 occurrences were considered (0.05% of the samples). In total, SELECT produced a list of 273 high-scoring motifs between 315 alterations. Pathway-level analysis of conditional selection The dependency motifs were summarized at pathway level by considering independently (i) the sum of motif scores between each pair of pathways, and (ii) the number of significant motifs. The significance of pathway-level interactions was empirically estimated by comparing the observed sum of motif scores and number of significant motifs to the null distribution obtained by randomly permuting the pathway annotation of the genes. The two metrics were first tested independently, and the two P values were then combined using Stouffer's method. Combined P values were then corrected with the Benjamini-Hochberg method. Corrected P values smaller than 0.25 were deemed to be significant. Curation of Clinical Trials The list of clinical trials for genes in pathways not represented in OncoKB was manually curated from ClinicalTrials.gov (http:// clinicaltrials.gov). Clinical trials with drug compounds targeting pathway members or which described pathway members in their inclusion or exclusion criteria are reported. Focus was given to ongoing clinical trials. A description of data retrieved from particular clinical trials is in the README worksheet of Table S5. If available, PubChem Compound IDs (https://pubchem.ncbi.nlm.nih.gov/) are given for drug compounds. QUANTIFICATION AND STATISTICAL ANALYSIS Quantitative and statistical methods are described above within the context of individual analyses in the Method Details section. DATA AND SOFTWARE AVAILABILITY The raw data, processed data and clinical data can be found at the legacy archive of the GDC (https://portal.gdc.cancer.gov/ legacy-archive/search/f) and the TCGA PanCancer Atlas publication page (https://gdc.cancer.gov/about-data/publications/ pancanatlas). The mutation data can be found here (https://gdc.cancer.gov/about-data/publications/mc3-2017). Data can also be visualized and downloaded using a dedicated section of the cBioPortal for Cancer Genomics (http://www. cbioportal.org/). ADDITIONAL RESOURCES Pathway diagrams were curated using PathwayMapper (Bahceci et al., 2017), a tool that allows visualization and design of pathway diagrams stylized as in classical TCGA publications. This tool is publicly available online at www.pathwaymapper.org. Our curated templates provided in Table S3 are accessible as pre-defined pathway diagrams that have been incorporated to the PahtwayMapper interface. PathwayMapper also acts as an interactive resource that allows to easily overlay user-inputed alteration frequencies on top of these predefined diagrams. Cell 773, 321-337.e1-e4, April 5, 2018 e4 Cell Supplemental Figures RTK/RAS pathway i urn urn II ERBB4 II MET~~|| PDGFR/ I FGFR1 || FGFR2 ||~FGFR3 | |~FGFR4 || KIT || IGF1R I RET |f R0S1 |[~ALK || FLT3 || NTRkT5| | JAK2 I CBL J |~ERRFI1 J J ABL1 |—»| S0S1 | | NF1 | |~RASA11 I PTPN11 J| HRÄS~|| NRAS~|| RIT1 I l RAS I ARAF It 1 | RAC1 | —» | MAPK1 | <— | MAP2K1 | [MAP2K2 MEK Proliferation Cell survival Translation Wnt pathway | wifi I RNF43 |i I ZNRF3 |- I SFRP1-5|-| [ WNT ligands] i I AXIN1/2 |-i i-rA^_|H(CTNNBi)r-l r |~DKK1-4 ptor C( i I GSK3B |h | WNT Dual Receptor Complex -1 Cell ||TCF7L1/2| proliferation TCF/LEF I TLE1-4 J Groucho Hippo pathway |~DCHS1/2 L I WWC1 I I CRB1/2 ]- I SAV1 If STK3/4 || MOB1A/B PTPN14 |CSNK1E/P| [tEAD2 I i Cell proliferation and differentiation Nrf2 pathway I KEAP1~|| CUL3 Oxidative Stress PI3K pathway I INPP4B I--|_ J_ j PTEN i PIK3CA/B —I PIK3R2 I- I PIK3R1/3 1 I PPP2R1A| —\ j AKT1/2/3| ■ -1 1 I STK11 |—► j TSC1/2 |—I j RHEB | J I RICTOR j mTORC2 I MTOR I I RPT0R I j mTORCI Cell growth Myc pathway | MAX MYCN |[JtfyCL i h MAX/MYC complex lr MXD1/3/4 I_ MAX/MGA complex Cell growth, proliferation, MXI1 j I MNT I MAX/MXD complex MLX I -1- I ' I MLXIP MLXIPL MLX/MONDO complex p53 pathway MDM2/4 | DNA replication stress Oncogenic I A™ I 1 TGFß pathway TGFß ligands Activin ligands 1 I TGFBR1/2J I ACVR2A/1B I SMAD2|| SMAD3 | f SMAD4j Proliferation, stem/progenitor phenotype Notch pathway I JAG2 j j ARRDC1 | | FBXW7 |<— | CUL1 | ,-, i ^ ,.- ,| I i ii ' ' i I |-»| HES-X U, NOV h I NOTCH11 PnOTCH2] I cteaved | MAML3 KAT2B _p >~ l-l-1 11 NOTCH311 NOTCH41 ~~* NOTCH ~~* | CREBBP I I EP300 I I—»I HEY-X [_ I CNTN6 I NQTCH3|| NOTCH4 t t I DNER 1| PSEN2 ] I CREBBP I I EP300 | T I NCQR1/2 j j SPEN j j KDM5A | Cell growth, —* Activation I Inhibition Fusions/Rearrangements I Part of complex Epigenetic silencing Copy number changes^_ MutationsHFGene | Alteration ttequendes Oncogene n<1% >1%Q>5%B>10% Tumor suppressor Q >!%□ >5%B >10% Figure S1. Curated Pathways Including Cross-Cross Pathway Interactions, Related to Figure 2 Same as Figure 2, but including cross-pathway interactions. o t> t> 33 £I3IIIIIIu->t>uu-'->!Tluu-'?IIIli GBM LGG IDHwt LGG IDHmut codel LGG IDHmut-non-codel UVM HNSC HPV+ HNSC HPV-THCA ACC PCPG THYM LUAD MESO LUSC BRCA LumA BRCA LumB BRCA Her2-enriched BRCA Basal BRCA Normal 5TE5 Squamous STES CIN STES EBV STES GS STES MSI STES POLE CRC MSI CRC POLE CRCGS CRC CIN LIHC CHOL PAAD KIRC KIRP KICH BLCA PR AD TGCTs. TGCT non-s OV UCEC CN high UCEC CN low UCEC MSI UCEC POLE ucs CESC Adeno CESC Squamous SKCM SARC DDLPS SARC LMS SARCMFS/UPS SARC other DLBC LAML o — ■> 5 - S O Ul _k -J . i] ^l — l [col I ro -g ^ £ ro co ro ; - ^ K -1 ro ro co j .IT NRF2 pathway i *• *• s JF I' TP53 pathway i 5 t * S 2 i £ a £ I S ' * - I » I »<» | 3 28: "ET1* => - ■ U> CO b | -»■ "1 Ph h ro w to - — _i ro ->■ po ro o o o Ho » S3 ti~F-B= | rui i n:i ■ N M Ji 1 ro — ■ I -I o o o o o □ a o o □ i * * * * * r 31 Z Z Z Z Z r 5 o S > S > i Cell cycle pathway o o o o mooooooo Niaaazzzz i Ul Ul », -o, ro ro 4s. 3 ro - Js. Ul Ul -1 "it a ■ ro co ro °> ro -1 CO -> ro to CD [ro CO — ro z CO -O. *. - - co .,-3) - S3 ro Ul ro - CO u s - - ** o> - CO CO era - Ul ro It - - - - CO CO - 8» - -■ ro 00 1-* CO - CO HI » ±g ro - -I _j - ro Ul Ul ro CO 5 co M ro 3 co CO ^> CO ro - $ o o OT co -1 -> CO CO ro ui 7^ ID u U Ul ro - It o> to co CO ,-CJ> w CTJ CO CO CT> CO .-co -1 Ul -J CO o ro - fo to CO 8 ^ ro - CO ro CO CD Lt ro ro co 3 ... CO E ► 3 4 973156 12 1 CO *. ro 4i 1- - CO -I ro ro -t> CO <*> CO Ul 0:1 ro ro ro ro ro CO CO CO o 0 0 0 i 1-Ul ro | ro ro ro 1 -ro s ^ w to ro ro CO *> CIl CO co CO cn CO -J CO CO w - - Ul ro Ul w w W m - ro W O - r° co co ro ro Ul ro co ro -■ Ul CO CO ro - o> ro ro ro era co ^ Ul ro - ro CO CO u. p 00 1. -oj ro CO CO —* CO ro - ro Ul 00 CO CO -1 ro -g - 5 ro -> t> S3 Cell Figure S2. Cell-Cycle, Wnt, p53, Nrf2 and PI3K Pathway Alterations, Related to Figure 5 Detailed heatmap of alteration frequencies in members of the Cell-cycle, WNT, TP53 and NRF2 pathways. Shades of red indicate activating event (mutation, amplification, activating fusion) and shades of blue indicate inactivating event (mutation, homozygous loss, inactivating fusion, epigenetic silencing). Color side bars show the fraction of samples affected by each type of somatic alteration (or a combination of them) for each pathway gene. Top color bars show the proportion of different types of alterations for each cancer subtype. ■9 CD Q. Notch pathway i!EB3i§Einm8glEiHnm!llh LGG IDHwt i LGG IDHmut-codel i LGG IDHmut-non-codel i ; HNSC HPV+1 HNSC HPV- ! thcaHI ACC PCPQ i STES Squamoua 5TESCIN STESEBV STES GS STES MSI STES POLE CRC MSI CRC POLE CRC GS CRC CIN LIHC CHOL 00 PAAD £ KIRC KIRP KICH BLCA | PRAD TGCT mm I noma a> TQCT non-aomlnoina 5> ■ I hi hi I |w| | |r»|-t| | ^ NMMM M 0> & -» u w « « UUUN jj S> OB £j^lOMMWrOIONIOIO MUytt | 3I00IS 0080 i UCEC CN high S UCEC CN low £ M-* UCEC MSI £ -1 - * £ « UCEC POLE g 00 o> ro £ H UCS g 10» CESCAdano g C ESC Squamoua S SKCM £j w SARCDDLPS ^ SARCLMS £ SARCMFS/UPS S SARC other 01 DLBC ^ LAML » » » ~SSSi M M *> M I 01 ui 01 01 loilrolrol I m -» 4* M -* u I to I —»■ I —*■ I —*■ I —^1 ro I —■■ I "1 iMiMi h r ■ :■. ".■ ::t. tii TGFB pathway „£££§§33 E> > > 9 bd n 2 S S 5 K 3 2 I 3 OB N 10 U -J "* M " ^ -J w u -a £8 IS - - |»|io| I r as "l u cn || Hippo pathway 1 CSNK1D CSNK1E □CHS1 1 RATI FAT2 FAT3 CRB3 CRB2 TAOK1 TAOK2 TAOK3 CRB1 NF2 WWC1 MOB1A MOB1B PTPN14 LATS2 LATS1 1 STK4 STK3 WWTR1 TEAD2 s •z 0 ro -1 ■ -1 s 8 cn -1 © OB -1 K > -1 ro ro CO » ■ 1 UI T u 4» -* W u ro 8 ■* ■* w ro a M —"-I—*-| ro ro | £ s — at ro 8 III —po ro -» -1 ro -1 | UI -1 UI I —»■ I —*■ I —1.1 —»■ I —ICJ1 u -1 -1 -1 a I |m -1 ■jj ^1 | |ro u -» -» -» -» — -» ■jj pjj u ■jjjjj1 2 ro en en — -1 -1 ro ijjj 0 1 I |ro| | ro ■jjj !l u a UI IS IS ■jj] i u 10 a ' CO ■-I o> a u ro ui ro ro u « 10 1_ ro 0 i ss 0 s s ■jjj 0 ro u ro jjj 00 | |m u jjj ro |-*| | A jjjj ■M u a, jj jjj -J ■* - * ■* U -1 [uT -»■ ro jj u -1 ro w ro 10 «> -1 -> ro *> jj UI - - jjjj ■ 0 ■ s ro w >■ BBBl 1 u ro -J -1 e ■O 1 CO OB UI u u ro -1 U -1 CO -1 —1 M -1 ~ S3 KS & VI w e OB ro 08 OB »| 1 1 ro | | |ui| | 10 ro *> ro ro jjjj 10 ro ro ro ro ro a — -* ro -t ro -* u u ro jjj a ro 10 01 ro u 10 ro ro - ro ro ->| | ro | | (0 ro ro jjjjjji a -* - -»■ ro ro j | O | 1 UI u OB u UI jjj W -* _ -* ■ .1 1. "1.11 H1 EU Cell Figure S3. Hippo, Myc, TGFp and Notch Pathway Alterations, Related to Figure 5 Detailed heatmap of alteration frequencies in members of the Hippo, Myc, TGFp and Notch pathways. Shades of red indicate activating event (mutation, amplification, activating fusion) and shades of blue indicate inactivating event (mutation, homozygous loss, inactivating fusion, epigenetic silencing). Color side bars show the fraction of samples affected by each type of somatic alteration (or a combination of them) for each pathway gene. Top color bars show the proportion of different types of alterations for each cancer subtype. Cell Co-occurrence MUT.FGFR3 MUT.KIT AMP.JAK2 AMP.FGFR2 FUSION.FGFR3 MUT.FGFR2 MUT.EGFR MUT.ERBB3 AMP.FGFR4 AMP.ERBB3 AMP.EGFR AMP.FGFR1 AMP.FGFR3 AMP.IGF1R AMP.ERBB2 MUT.ERBB2 ^ ^ < cm ^ i " 5 f. cm z ^ <*> Z £ ^ 77 CD 7- -> Q. Q. G MUT.TP53 MUT.ATM AMP.MDM4 AMP.MDM2 MUT.RB1 MUT.CDKN2A EPISIL.CDKN2A DEL.RB1 DEL.CDKN2A AMP.E2F3 AMP.CDK6 AMP.CDK4 AMP.CCNE1 AMP.CCND1 HUD y HUH ■ cd co < t- < t- cm cm co cm cm co ^ ^ s 2 < < PI3K/Akt patthway cell cycle p53 pathway SELECT score Co-occurence | Mutual Exclusivity * Significant MUT.TP53 ■ MUT.ATM ■ AMP.MDM4 ■ DEL.ATM ■ DEL.TP53 ■ MUT.CHEK2 ■ AMP.MDM2 ■ z z 11 LU 3. Q. < ^ cm ^ 5 h o: o: n ^ o n ^ < ^ * -E cm < 2 Z < I" tQ Ü UJ Ü ^ ^ o I— co co Q. ^ Q_ ^ ^ D- MUT.TAOK1 MUT.TAOK2 MUT.FAT1 1 MUT.NF2 sz ro MUT.FAT3 Q. o MUT.FAT4 Q. Q. MUT.FAT2 I DEL.FAT1 MUT.DCHS2 AMP.YAP1 AMP.NFE2L2 MUT.KEAP1 cell cycle □ 2 1 2 PI3K/Akt patthway s 2 1 =1 2 p53 pathway Figure S4. Additional Results for Conditional Selection, Related to Figure 6 (A-E) Mutual exclusivity (in purple) and Co-occurrence (in green) between alterations in (A) PI3K and RTK pathways, (B) Cell-cycle and p53 pathways, (C) p53 and PI3K pathways, (D) p53 and Hippo pathways, and (E) cell-cycle and Nrf2 pathways.