File size: 32,521 Bytes
c633a0a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 |
[
{
"id": 0,
"page": 5,
"bounding_box": [
72.0,
82.7750244140625,
575.864990234375,
667.2822723388672
],
"latex_content": "\\begin{table}[!hp]\n\\setcounter{table}{0}\n\\makeatletter \n\\renewcommand{\\thetable}{\\@arabic\\c@table} \n\\makeatother\n\n\\setlength{\\LTcapwidth}{\\textwidth}\n\\renewcommand{\\arraystretch}{1.5}\n\\begin{longtable}{|p{1in}|p{4in}|p{1.5in}|}\n\\caption{Listing of the 18 attractor signatures}\\\\\n\\hline\n\\textbf{Name} & \\textbf{Top members} & \\textbf{Comments}\\\\\n\\hline\n\\endfirsthead\n\n\\multicolumn{3}{c}\n{{\\bfseries \\tablename \\thetable{} -- continued from previous page}} \\\\\n\\hline \n\\textbf{Name} & \\textbf{Top members} & \\textbf{Comments}\\\\ \n\\hline \n\\endhead\n\n\\hline \\multicolumn{3}{r}{{Continued on next page...}} \\\\\n\\endfoot\n\n\\hline \\hline\n\\endlastfoot\n\n\\hline\n\\multicolumn{3}{|l|}{\\textbf{mRNA}}\\\\\n\\hline\n\\textbf{LYM}\t&\t\\textit{SASH3, CD53, NCKAP1L, LCP2, IL10RA, PTPRC, EVI2B, BIN2, WAS, HAVCR2}\t&\tlymphocyte infiltration\\\\\n\\hline\n\\textbf{CIN}\t&\t\\textit{TPX2, KIF4A, KIFC1, NCAPG, BUB1, NCAPH, CDCA5, KIF2C, PLK1, CENPA}\t&\tmitotic chromosomal instability\\\\\n\\hline\n\\textbf{MES}\t&\t\\textit{COL3A1, COL5A2, COL1A2, THBS2, COL5A1, VCAN, COL6A3, SPARC, AEBP1, FBN1}\t&\tmesenchymal transition \\\\\n\\hline\n\\textbf{END}\t&\t\\textit{CDH5, ROBO4, CXorf36, CD34, CLEC14A, ARHGEF15, CD93, LDB2, ELTD1, MYCT1} \t&\tendothelial markers\\\\\n\\hline\n\\textbf{``\\textit{AHSA2}''}\t&\t\\textit{AHSA2, LOC91316, PILRB, ZNF767, TTLL3, CCNL2, PABPC1L, LENG8, CHKB CPT1B, SEC31B}\t&\\\\\n\\hline\n\\textbf{IFIT}\t&\t\\textit{IFIT3, MX1, OAS2, RSAD2, CMPK2, IFIT1, IFI44L, IFI44, IFI6, OAS1}\t&\tinterferon-induced\\\\\n\\hline\n\\textbf{``\\textit{WDR38}''}\t&\t\\textit{WDR38, YSK4, ROPN1L, C1orf194, MORN5, WDR16, RSPH4A, FAM183A, ZMYND10, DNAI1}\t&\t\\\\\n\\hline\n\\multicolumn{3}{|l|}{\\textbf{Genomically co-localized mRNA}} \\\\\n\\hline\n\\textbf{MHC Class II}\t&\t\\textit{HLA-DPA1, HLA-DRA, HLA-DPB1, HLA-DRB1, HLA-DMA, HLA-DMB, HLA-DOA, HLA-DQA1, HLA-DRB5}\t&\tstrongly associated with LYM\\\\\n\\hline\n\\textbf{GIMAP cluster}\t&\t\\textit{GIMAP4, GIMAP7, GIMAP6, GIMAP5, GIMAP8, GIMAP1}\t&\tstrongly associated with LYM\\\\\n\\hline\n\\textbf{Chr8q24.3 amplicon}\t&\t\\textit{SHARPIN, HSF1, TIGD5, GPR172A, ZC3H3, EXOSC4, SCRIB, CYHR1, MAF1, PUF60} \t&\tmost prominent Pan-Cancer amplicon\\\\\n\\hline\n\\multicolumn{3}{|l|}{\\textbf{microRNA}} \\\\\n\\hline\n\\textbf{\\textit{DLK1}-\\textit{DIO3} RNA cluster}\t&\tmir-127, mir-134, mir-379, mir-409, mir-382, mir-758, mir-381, mir-370, mir-654, mir-431\t&\tincludes \\textit{MEG3} long noncoding RNA; associated with MES\\\\\n\\hline\n\\textbf{``mir-509''}\t&\tmir-509, mir-514, mir-508\t&\t\\\\\n\\hline\n\\textbf{``mir-144''}\t&\tmir-144, mir-451, mir-486\t&\tassociated with erythropoiesis\\\\\n\\hline\n\\multicolumn{3}{|l|}{\\textbf{Methylation}}\\\\\n\\hline\n\\textbf{``RMND1''}\t&\tRMND1-6-151814639, MAP3K7-6-91353911, DNAAF1-16-82735714, PTRH2-17-55139429, ZNF143-11-9439170, cg03627896\n, TAMM41-3-11863582, CDK5-7-150385869, OTUB1-11-63510174, AATF-17-32380976\t&\t\\\\\n\\hline\n\\textbf{M+}\t&\tcg13928306, MTMR11-1-148175405, cg27324619, TNKS1BP1-11-56846646, C11orf52-11-111294703, IL17RC-3-9934128, cg24765079, ERBB3-12-54759072, IL22RA1-1-24342151, C11orf52-11-111294903\t&\tmethylated in infiltrating lymphocytes\\\\\n\\hline\n\\textbf{M-}\t&\tBIN2-12-50003941, PTPRCAP-11-66961771, TNFAIP8L2-1-149395922, IGFLR1-19-40925164, FAM113B-12-45896487, CD6-11-60495754, KLHL6-3-184755939, PTPN7-1-200396189, FAM78A-9-133141340, ACAP1-17-7180947\t&\tUnmethylated in infiltrating lymphocytes, may be causal to the expression of some of the genes of the LYM signature\\\\\n\\hline\n\\multicolumn{3}{|l|}{\\textbf{Protein activity}} \\\\\n\\hline\n\\textbf{``c-Met''}\t&\tc-Met, Snail, PARP\\_cleaved, Caspase-8, ERCC1, Rb\t&\tRelated to apoptosis\\\\\n\\hline\n\\textbf{``Akt''}\t&\tAkt, Tuberin, STAT5A\t&\\\\\n\\end{longtable}\n\n\\begin{figure}[!p]\n\\fbox{\n\\begin{minipage}{6.5in}\n\\includegraphics[width=\\textwidth]{figure2.png}\n\\caption{\nScatter plots connecting the LYM, M+ and M- metagenes in 12 cancer types. Each dot represents a cancer sample. The horizontal and vertical axes measure the average methylation values of the two methylation signatures, M- and M+, while the value of the expression of the LYM metagene is color-coded. In all three cases, the metagene is defined by the average of the top ranked genes as described in \\textbf{Table ~\\ref{tab:tabS1}}. \n}\n\\label{fig:fig2}\n\\end{minipage}\n}\n\\end{figure}\n\n\\subsection{Lymphocyte infiltration: LYM mRNA signature; M+\\hspace{1pt} methylation signature; M-\\hspace{1pt}\u00a0methylation signature}\nThese three signatures are related to tumor infiltration by lymphocytes. We list \nthem together because they are strongly interrelated (\\textbf{Fig. ~\\ref{fig:fig2}}) even though each \nof the three was independently derived using an unsupervised computational method. \nThe presence of LYM is accompanied by the presence of M+ and the absence of M- in \nall solid cancer types, suggesting that the three signatures reflect the same \nbiomolecular event, which appears to be the infiltration of immune cells in tumor \ntissue. Indeed, there is remarkable similarity (\\textbf{Fig. ~\\ref{fig:fig3}}) between the LYM signature \nand the ``immune score'' of the ESTIMATE tumor purity computational tool (\\url{http://ibl.mdanderson.org/estimate}). \nThe values of the M+ methylation signature are also remarkably similar to those of the methylation-based \n``leukocyte percentage'' estimation \\cite{huiNature} (available under Synapse ID syn1809222).\n\n\\begin{figure}[!p]\n\\fbox{\n\\begin{minipage}{6.5in}\n\\includegraphics[width=\\textwidth]{figure3.png}\n\\caption{\nScatter plots demonstrating the pan-cancer similarity of the value of the LYM metagene with the immune score of the ESTIMATE tumor purity computational tool (http://ibl.mdanderson.org/estimate) measuring immune cell infiltration. Each dot represents a cancer sample. The horizontal axis measures the expression value of the LYM metagene and the vertical axis measures the ESTIMATE immune score of infiltration. Note that the ESTIMATE did not provide scores for rectum cancer, and the estimation of immune cell infiltration is not applicable in leukemia.\n}\n\\label{fig:fig3}\n\\end{minipage}\n}\n\\end{figure}\n\nWe had previously found\\cite{billCancerInfo} all three LYM, M+ and M- signatures from their association with \nthe expression of miR-142. We have now confirmed this association with miR-142 in the pancan12 data sets, \nand we found that miR-150 and miR-155 are also strongly associated with the LYM signature. \nWe had also previously independently identified the LYM signature as an attractor metagene\\cite{mePLoS}, and \nused it in the winning model of the Sage Bionetworks Breast Cancer Prognosis Challenge\\cite{meSTM}. \nSpecifically the LYM signature is strongly associated with improved prognosis in ER-negative breast cancers, \nand this fact also provides an explanation for the relatively better prognosis in medullary, \ncompared with other types of high-grade breast cancers. \n\nThe interrelationship of the LYM, M+ and M- signatures, as shown in \\textbf{Fig. ~\\ref{fig:fig2}}, appears to be a consequence \nof the presence of different subclasses of cells (as opposed to being a methylation switch inside \nthe same cell), consistent with their assumed role of measuring the extent of lymphocyte infiltration \nin the tumor. In other words, the M+ methylation sites, normally unmethylated, are largely methylated \nin the infiltrating leukocytes; and the M- methylation sites, normally methylated, are largely \nunmethylated in the infiltrating leukocytes. Consistently, many of the genes methylated by the M- \nsignature are identical to those of LYM (six among the 27 genes of the M- signature (\\textit{BIN2}, \\textit{TNFAIP8L2}, \n\\textit{ACAP1}, \\textit{NCKAP1L}, \\textit{FAM78A}, \\textit{PTPN7}) listed in Table S1 are also among the 168 \ngenes listed in the LYM attractor metagene \n($P < 9.21\\times10^{-7}$ based on Fisher\u2019s exact test). The observed significant overlap in the gene sets and \nthe negative association between gene expression in LYM and DNA methylation in M- are consistent with the notion \nthat the absence of DNA methylation is permissive for gene expression, suggesting that the expression of the LYM\nsignature in the infiltrating lymphocytes may be facilitated in part by the hypomethylation of the M- signature.\n\n\n\nThe sharp definition of the LYM signature (being a Pan-Cancer attractor signature pointing to few genes \nat the core of coexpression) provides strong hints about the precise nature of this leukocyte infiltration. \nSpecifically, the membership of the top-ranked genes (\\textit{SASH3}, \\textit{CD53}, \\textit{NCKAP1L}, \\textit{LCP2}, \n\\textit{IL10RA}, \\textit{PTPRC}, \\textit{EVI2B}, \\textit{BIN2}, \\textit{WAS}, \\textit{HAVCR2}, \\ldots) \npoint to a specific type of lymphocytes. We have speculated\\cite{mePLoS} that these infiltrating lymphocytes \nare T cells having undergone a particular type of co-stimulation providing hypotheses for related adoptive transfer therapy.\n\nTwo proteins strongly associated with the LYM signature are two tyrosine kinases: Lck (lymphocyte-specific protein tyrosine kinase) \nand Syk (spleen tyrosine kinase).\n\n\\subsection{CIN (mitotic chromosomal instability) mRNA signature}\nThis signature is related to mitotic chromosomal instability. It is similar to numerous known ``proliferation'' signatures, \nbut its sharp definition as an attractor metagene specifically points to the kinetochore-microtubule interface and \nassociated kinesins. Comparison with similar mitotic signatures in normal cells may help pinpoint driver genes for \nmalignant chromosomal instability. The signature is strongly associated with tumor grade as well as poor prognosis in many, \nif not all, cancer types.\n\nTwo proteins strongly associated with the CIN signature are Cyclin B1 and CDK1. Consistently, it is known that the cyclin \nB1-Cdk1 complex of cyclin-dependent kinase 1 is involved in the early events of mitosis, and that nuclear cyclin B1 protein \nmay induce chromosomal instability and enhance the aggressiveness of the carcinoma cells\\cite{bibi7}. \n\n\\subsection{MES (mesenchymal transition) mRNA signature}\nThis signature is related to mesenchymal transition and invasiveness of cancer cells. It is similar to numerous \n``stromal'' or ``mesenchymal'' signatures; however there is evidence\\cite{dimitrisBMCCancer} that many among the genes of the \nsignature are largely produced by transdifferentiated cancer cells. We hypothesize that such cells, known to \nassume the duties of cancer-associated fibroblasts in some tumors\\cite{hanahan2011}, may have become indistinguishable, \neven using laser capture microdissection, from stromal fibroblasts. We had originally identified the MES signature \nfrom its association with tumor stage\\cite{hoonBMCMedGenomics}; specifically the signature appears only after a particular cancer \ntype-specific tumor stage threshold has been reached.\n\nThe values of the MES signature are remarkably similar to the ``stromal score'' of the ESTIMATE tumor purity computational tool \n(\\url{http://ibl.mdanderson.org/estimate}) measuring fibroblast infiltration. Based on our previous reasoning, however, \nwe believe that this interpretation may not be fully accurate, and that it will be important to find out to what extent \nsome of the cells expressing some of these mesenchymal markers may actually be transdifferentiated cancer cells, and \nwhether the estimated tumor purity may be affected by other types of normal cells instead of stromal fibroblasts.\n\nThe co-regulated microRNAs most strongly associated with the MES signature are miR-199a, miR-199b, and miR-214. The \\textit{DLK1}-\n\\textit{DIO3} RNA cluster attractor signature, described later, is also strongly associated with MES.\n\nThe protein most strongly associated with the MES signature is Fibronectin.\n\n\\subsection{END (endothelial marker) mRNA signature}\nThis is a novel angiogenesis-associated attractor signature. Nearly all the top-ranked genes (\\textbf{Table 1}) are \nendothelial markers. The top gene, \\textit{CDH5}, codes for VE-cadherin, which is known to be involved in a pathway suppressing \nangiogenic sprouting\\cite{bib9}. The second gene, \\textit{ROBO4}, is known to inhibit VEGF-induced pathologic angiogenesis and endothelial \nhyperpermeability\\cite{bib10}. Consistently, the END attractor metagene appears to be protective and anti-angiogenic, stabilizing the \nvascular network. For example, 22 out of the 27 genes of the END attractor are among the 265 genes included in File S2 of a recent \nstudy\\cite{bib11} of renal cell carcinoma ($P < 8.4\\times10^{-38}$ based on Fisher\u2019s exact test) as most associated with patients\u2019 \nsurvival. These good-prognosis genes were intermixed in the same file with many poor-prognosis genes of the CIN attractor, \nsuggesting that the CIN and END attractor metagenes are two of the most prognostic features in renal cell carcinoma. \n\nInterestingly, the MES and END attractor metagenes are positively associated with each other (\\textbf{Fig. ~\\ref{fig:fig4}}), in the sense \nthat overexpression of the END signature tends to imply overexpression of the MES signature and vice-versa. This is consistent \nwith mutual exclusivity between angiogenesis and invasiveness and with related findings\\cite{bib12} that VEGF inhibits tumor cell \ninvasion and mesenchymal transition, while antiangiogenic therapy is associated with increased invasiveness\\cite{bib13}. It may also \nexplain the paradoxical protective nature of signatures related to the MES attractor metagene in invasive breast cancers\\cite{bib14}.\n\n\\begin{figure}[!p]\n\\fbox{\n\\begin{minipage}{6.5in}\n\\includegraphics[width=\\textwidth]{figure4.png}\n\\caption{\nScatter plots demonstrating the association between MES and END attractor metagenes. The horizontal and vertical axes measure the values of the MES and END signatures. The two signatures have positive correlation, although this association is not sufficiently strong to merge the two attractors into one. This association suggests that the invasive MES signature and the antiangiogenic END signature tend to be present simultaneously.\n}\n\\label{fig:fig4}\n\\end{minipage}\n}\n\\end{figure}\n\n\n\\subsection{``AHSA2'' mRNA signature}\nWe do not yet know what this signature represents. We observed that several noncoding RNAs (e.g. NCRNA00105, NCRNA00201) \nare in relatively high-ranked positions among its members.\n\n\\subsection{IFIT (interferon-induced) mRNA signature}\nThe members of this signature are interferon-induced. For example, we observed large enrichment of the genes of the \nsignature among those upregulated by IFN-$\\alpha$ in the side population (SP) of ovarian cancer cells\\cite{bib15} from the \nlist provided in Supplementary Table S4 of that paper, in which the authors concluded that tumors bearing large SP numbers \ncould be particularly sensitive to IFN-$\\alpha$ treatment. \n\n\\subsection{``WDR38'' mRNA signature}\nWe do not know what this signature represents, except that we had found one of its key members, gene \\textit{ZMYND10}, to be \nprotective and associated with estrogen receptor expression in breast cancer.\n\n\\subsection{MHC Class II genomically co-localized mRNA signature}\nWe found this signature using the genomically co-localized version of the algorithm. It is very highly correlated with LYM.\n\n\\subsection{GIMAP genomically co-localized mRNA signature}\nAs above, we found this signature using the genomically co-localized version of the algorithm. It is also very highly correlated with LYM.\n\n\\subsection{Chr8q24.3 amplicon mRNA signature}\nThis is the strongest pan-cancer amplicon signature. It was previously found predictive of early relapse in ER-positive breast cancers\\cite{bib16}.\n\n\\subsection{``RMND1'' methylation signature}\nWe do not yet know what the comethylation of the sites of this signature signifies.\n\n\\begin{figure}[!t]\n\\fbox{\n\\begin{minipage}{6.5in}\n\\includegraphics[width=\\textwidth]{figure5.png}\n\\caption{\nThe \\textit{DLK1}-\\textit{DIO3} cluster of noncoding RNAs. Shown is a screen capture from the UCSC Genome Browser (\\url{http://genome.ucsc.edu}). The cluster of imprinted genes delineated by the \\textit{DLK1} and \\textit{DIO3} genes (outside the shown region) is located on chromosome 14. We found that the corresponding pan-cancer attractor signature does not contain any paternally inherited protein-coding genes. It does contain the numerous noncoding RNA genes expressed from the maternally inherited homolog, including the \\textit{MEG3} long noncoding RNA gene.\n}\n\\label{fig:fig5}\n\\end{minipage}\n}\n\\end{figure}\n\n\\subsection{\\textit{DLK1}-\\textit{DIO3} RNA cluster signature}\nThis is the strongest pan-cancer multi-microRNA coexpression signature. It consists of numerous noncoding RNAs within the \n\\textit{DLK1}-\\textit{DIO3} imprinted genomic region of chr14q32. \\textbf{Fig. ~\\ref{fig:fig5}} shows a screen capture of the genomic region \nfrom the UCSC Genome Browser (\\url{http://genome.ucsc.edu/}). We confirmed that the coexpression signature also includes the \n\\textit{MEG3} long noncoding RNA located at the upstream end of the region. It may also include numerous small nuclear \nRNAs at the central region, but there were no associated probe sets to confirm the coexpression. We found that this ncRNA \nsignature is associated with the MES (mesenchymal transition) mRNA signature. For example, the ranked list of mRNAs most \nassociated with the \\textit{DLK1}-\\textit{DIO3} ncRNA signature starts from \\textit{POSTN}, \\textit{PCOLCE}, \\textit{COL5A2}, \\textit{COL1A2}, \n\\textit{GLT8D2}, \\textit{COL5A1}, \\textit{SFRP2}, and \\textit{FAP}.\n\nExpression of the imprinted \\textit{DLK1}-\\textit{DIO3} ncRNA cluster is believed to be vital for the development potential of \nembryonic stem cells\\cite{bib17}, consistent with the hypothesis\\cite{bib18} that mesenchymal transition in cancer reactivates \nembryonic developmental programs and makes cancer cells invasive and stem-like. The \\textit{DLK1}-\\textit{DIO3} ncRNA signature \nwas also found to define a stem-like subtype of hepatocellular carcinoma associated with poor survival\\cite{bib19}. The details \nof the regulation mechanism for this ncRNA cluster coexpression in the \\textit{DLK1}-\\textit{DIO3} region are unclear.\n\n\\subsection{``miR-509/miR-514/miR-508'' microRNA signature}\nThese three microRNAs are co-localized at chrXq27.3. We do not know what this signature signifies.\n\n\\subsection{``miR-144/miR-451/miR-486'' microRNA signature}\nThis is a three-microRNA signature related to erythropoiesis. The first two genes are located in the bicistronic microRNA \nlocus miR-144/451, highly expressed during erythrocyte development\\cite{bib20}. The mRNAs most associated to this microRNA \nsignature are hemoglobin-related: \\textit{HBB}, \\textit{HBA1}, \\textit{HBA2} and \\textit{ALAS2}. The protein most associated \nwith this signature is HER3. These three microRNAs were identified as promising biomarkers for detection of esophageal cancer.\n\n\\subsection{c-Met/Snail/PARP\\_cleaved/Caspase-8/ERCC1/Rb protein activity signature}\nThis protein coexpression signature appears to combine the contribution of several pathways and we hope that a plausible \nand useful biological ``story'' will be developed based on the simultaneous activity of all these six proteins in some \ncancer samples. We note that each of these proteins\\cite{bib21, bib22, bib23, bib24, bib25, bib26} has been related in various ways with resistance to chemotherapy \nor apoptosis.\n\n\\subsection{Akt/Tuberin/STAT5A protein activity signature}\nWe do not know what the coexpression of Akt, Tuberin, STAT5A proteins represents in cancer. It is known, however, that \nlow levels of STAT5A protein in breast cancer are associated with tumor progression and unfavorable clinical outcomes\\cite{bib27}.\n\n\\section*{DISCUSSION}\nThe Pan-Cancer nature (\\textbf{Fig. ~\\ref{fig:figS1}}) of each of the signatures described in this paper suggests that they represent \nimportant biomolecular events. A reasonable concern is whether some of these ``pan-cancer'' signatures may instead reflect \nfundamental normal ``pan-tissue'' biological mechanisms. Even if this is true for some of these signatures, this does not \nexclude the possibility that they are aberrant and play important roles in some cancer samples. Furthermore, this provides \nthe opportunity to compare similar signatures in normal vs. malignant tissues to pinpoint potential cancer-specific genes. \n\nBecause of its exhaustive search starting from all potential ``seeds'' in all data sets from twelve different cancer types, \nour iterative data mining algorithm is guaranteed to have identified all pan-cancer molecular signatures involving simultaneous \npresence of a large number of coordinately expressed genes, proteins, or comethylated sites. We hope that these signatures are \nfurther scrutinized by the medical research community for the purpose of developing potential diagnostic, predictive, and \neventually therapeutic products applicable in multiple cancers. \n\n\\section*{ACKNOWLEDGEMENTS}\nWe are thankful to Hanina Hibshoosh, Chris Miller and Gordon Mills for helpful discussions, \nwhich contributed to improved interpretation of the signatures disclosed in this paper.\n\n\\section*{Accessibility} All figures in this paper, including supplementary figures and tables, as well as the files of generated attractor molecular signatures, are available in Synapse under ID syn1686966. \n\n\\section*{Data description and availability} The data sets of TCGA pancan12 freeze 4.7 used to derive the results of this paper are described and are available under Synapse ID syn300013 with doi:10.7303/syn300013. The twelve cancer types are bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), colon adenocarcinoma (COAD), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), acute myeloid leukemia (LAML), lung adenocarcinoma (LUAD) , lung squamous cell carcinoma (LUSC), ovarian serous cystadenocarcinoma (OV), rectum adenocarcinoma (READ), and uterine corpus endometrioid carcinoma (UCEC).\n\n\\pagebreak\n%% Put the bibliography here, most people will use BiBTeX in\n%% which case the environment below should be replaced with\n%% the \\bibliography{} command.\n\n\\begin{thebibliography}{1}\n\\bibitem{hanahan2000} Hanahan, D. \\& Weinberg, R.A. The hallmarks of cancer. Cell 100, 57-70 (2000).\n\\bibitem{hanahan2011} Hanahan, D. \\& Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646-74 (2011). \n\\bibitem{mePLoS} Cheng, W.Y., Ou Yang, T.H. \\& Anastassiou, D. Biomolecular events in cancer revealed by attractor metagenes. PLoS Comput Biol 9, e1002920 (2013).\n\\bibitem{meSTM} Cheng, W.Y., Ou Yang, T.H. \\& Anastassiou, D. Development of a prognostic model for breast cancer survival in an open challenge environment. Sci Transl Med 5, 181ra50 (2013).\n\\bibitem{bib5} McCarthy, N. Prognostic models: Rising to the challenge. Nat Rev Cancer, doi:10.1038/nrc3530 (2013).\n\\bibitem{huiNature} Shen, H. \\textit{et al.} Comprehensive Cross-Cancer Comparison of DNA Methylation Profiles, submitted. (2013).\n\\bibitem{billCancerInfo} Andreopoulos, B. \\& Anastassiou, D. Integrated Analysis Reveals hsa-miR-142 as a Representative of a Lymphocyte-Specific Gene Expression and Methylation Signature. Cancer Inform 11, 61-75 (2012).\n\\bibitem{bibi7} Suzuki, T. \\textit{et al.} Nuclear cyclin B1 in human breast carcinoma as a potent prognostic factor. Cancer Sci 98, 644-51 (2007).\n\\bibitem{dimitrisBMCCancer} Anastassiou, D. \\textit{et al.} Human cancer cells express Slug-based epithelial-mesenchymal transition gene expression signature obtained in vivo. BMC Cancer 11, 529 (2011).\n\\bibitem{hoonBMCMedGenomics} Kim, H., Watkinson, J., Varadan, V. \\& Anastassiou, D. Multi-cancer computational analysis reveals invasion-associated variant of desmoplastic reaction involving INHBA, THBS2 and COL11A1. BMC Med Genomics 3, 51 (2010).\n\\bibitem{bib9} Abraham, S. \\textit{et al.} VE-Cadherin-mediated cell-cell interaction suppresses sprouting via signaling to MLC2 phosphorylation. Curr Biol 19, 668-74 (2009).\n\\bibitem{bib10} Jones, C.A. \\textit{et al.} Robo4 stabilizes the vascular network by inhibiting pathologic angiogenesis and endothelial hyperpermeability. Nat Med 14, 448-53 (2008).\n\\bibitem{bib11} Wozniak, M.B. \\textit{et al.} Integrative genome-wide gene expression profiling of clear cell renal cell carcinoma in Czech Republic and in the United States. PLoS One 8, e57886 (2013).\n\\bibitem{bib12} Lu, K.V. \\textit{et al.} VEGF inhibits tumor cell invasion and mesenchymal transition through a MET/VEGFR2 complex. Cancer Cell 22, 21-35 (2012).\n\\bibitem{bib13} Paez-Ribes, M. \\textit{et al}. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell 15, 220-31 (2009).\n\\bibitem{bib14} Beck, A.H., Espinosa, I., Gilks, C.B., van de Rijn, M. \\& West, R.B. The fibromatosis signature defines a robust stromal response in breast carcinoma. Lab Invest 88, 591-601 (2008).\n\\bibitem{bib15} Moserle, L. \\textit{et al}. The side population of ovarian cancer cells is a primary target of IFN-alpha antitumor effects. Cancer Res 68, 5658-68 (2008).\n\\bibitem{bib16} Bilal, E. \\textit{et al}. Amplified Loci on Chromosomes 8 and 17 Predict Early Relapse in ER-Positive Breast Cancers. PLoS One 7, e38575 (2012).\n\\bibitem{bib17} Stadtfeld, M. \\textit{et al}. Aberrant silencing of imprinted genes on chromosome 12qF1 in mouse induced pluripotent stem cells. Nature 465, 175-81 (2010).\n\\bibitem{bib18} Mani, S.A. \\textit{et al}. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704-15 (2008).\n\\bibitem{bib19} Luk, J.M. \\textit{et al}. DLK1-DIO3 genomic imprinted microRNA cluster at 14q32.2 defines a stemlike subtype of hepatocellular carcinoma associated with poor survival. J Biol Chem 286, 30706-13 (2011).\n\\bibitem{bib20} Yu, D. \\textit{et al}. miR-451 protects against erythroid oxidant stress by repressing 14-3-3zeta. Genes Dev 24, 1620-33 (2010).\n\\bibitem{bib21} Tang, M.K., Zhou, H.Y., Yam, J.W. \\& Wong, A.S. c-Met overexpression contributes to the acquired apoptotic resistance of nonadherent ovarian cancer cells through a cross talk mediated by phosphatidylinositol 3-kinase and extracellular signal-regulated kinase 1/2. Neoplasia 12, 128-38 (2010).\n\\bibitem{bib22} Haslehurst, A.M. \\textit{et al}. EMT transcription factors snail and slug directly contribute to cisplatin resistance in ovarian cancer. BMC Cancer 12, 91 (2012).\n\\bibitem{bib23} D\u2019Amours, D., Sallmann, F.R., Dixit, V.M. \\& Poirier, G.G. Gain-of-function of poly(ADP-ribose) polymerase-1 upon cleavage by apoptotic proteases: implications for apoptosis. J Cell Sci 114, 3771-8 (2001).\n\\bibitem{bib24} Kim, P.K., Mahidhara, R. \\& Seol, D.W. The role of caspase-8 in resistance to cancer chemotherapy. Drug Resist Updat 4, 293-6 (2001).\n\\bibitem{bib25} Olaussen, K.A. \\textit{et al}. DNA repair by ERCC1 in non-small-cell lung cancer and cisplatin-based adjuvant chemotherapy. N Engl J Med 355, 983-91 (2006).\n\\bibitem{bib26} Collard, T.J. \\textit{et al}. The retinoblastoma protein (Rb) as an anti-apoptotic factor: expression of Rb is required for the anti-apoptotic function of BAG-1 protein in colorectal tumour cells. Cell Death Dis 3, e408 (2012).\n\\bibitem{bib27} Peck, A.R. \\textit{et al}. Low levels of Stat5a protein in breast cancer are associated with tumor progression and unfavorable clinical outcomes. Breast Cancer Res 14, R130 (2012).\n\n\n\\end{thebibliography}\n\n\\pagebreak\n\\begin{methods}\n\\subsection{Data normalization}\nThe data platform for each cancer types and its corresponding Synapse ID is given below. \n\n\\begin{table}[!hf]\n\\renewcommand{\\arraystretch}{1}\n\\hspace*{-0.5in}\n\\begin{threeparttable}\n\\begin{tabular}{ |l|c|c|c|c|}\n\\hline\n\\textbf{Molecular profile} & mRNA & Protein & miRNA & DNA methylation \\\\ \n\\hline\n\\multirow{3}{*}{\\textbf{Platform}} & & Reverse phase protein & & Infinium \\\\\n & Illumina HiSeq & lysate microarray & Illumina HiSeq & HumanMethylation27 \\\\\n & & (RPPA) & & BeadChip \\\\\n\\hline\n\\textbf{Cancer type} & \\multicolumn{4}{|c|}{\\textbf{Synapse ID}} \\\\\n\\hline\nBLCA&syn1571504&syn1681048&syn1571494&syn1889358\\tnote{*}\\\\ \nBRCA&syn417812&syn1571267&syn395575&syn411485\\\\ \nCOAD&syn1446197&syn416772&syn464211&syn411993\\\\ \nGBM&syn1446214&syn416777&NA&syn412284\\\\ \nHNSC&syn1571420&syn1571409&syn1571411&syn1889356\\tnote{*}\\\\ \nKIRC&syn417925&syn416783&syn395617&syn412701\\\\ \nLAML&syn1681084&NA&syn1571533&syn1571536\\\\ \nLUAD&syn1571468&syn1571446&syn1571453&syn1571458\\\\ \nLUSC&syn418033&syn1367036&syn395691&syn415758\\\\ \nOV&syn1446264&syn416789&syn1356544&syn415945\\\\ \nREAD&syn1446276&syn416795&syn464222&syn416194\\\\ \nUCEC&syn1446289&syn416800&syn395720&syn416204\\\\ \n\\hline\n\\end{tabular}\n\\begin{tablenotes}\n \\item[*] The data sets were extracted from HumanMethylation450 BeadChip.\n\\end{tablenotes}\n\\end{threeparttable}\n\\end{table}",
"extracted_content": [
[
"Name",
"Top members",
"Comments"
],
[
"END",
"CDH5, ROBO4, CXorf36, CD34, CLEC14A, ARHGEF15,\nCD93, LDB2, ELTD1, MYCT1",
"endothelial markers"
],
[
"\u201cAHSA2\u201d",
"AHSA2, LOC91316, PILRB, ZNF767, TTLL3, CCNL2,\nPABPC1L, LENG8, CHKB CPT1B, SEC31B",
""
],
[
"IFIT",
"IFIT3, MX1, OAS2, RSAD2, CMPK2, IFIT1, IFI44L, IFI44,\nIFI6, OAS1",
"interferon-induced"
],
[
"\u201cWDR38\u201d",
"WDR38, YSK4, ROPN1L, C1orf194, MORN5, WDR16,\nRSPH4A, FAM183A, ZMYND10, DNAI1",
""
],
[
"Genomically co-localized mRNA",
null,
null
],
[
"MHC Class II",
"HLA-DPA1, HLA-DRA, HLA-DPB1, HLA-DRB1, HLA-\nDMA, HLA-DMB, HLA-DOA, HLA-DQA1, HLA-DRB5",
"strongly associated\nwith LYM"
],
[
"GIMAP clus-\nter",
"GIMAP4, GIMAP7, GIMAP6, GIMAP5, GIMAP8, GIMAP1",
"strongly associated\nwith LYM"
],
[
"Chr8q24.3\namplicon",
"SHARPIN, HSF1, TIGD5, GPR172A, ZC3H3, EXOSC4,\nSCRIB, CYHR1, MAF1, PUF60",
"most prominent Pan-\nCancer amplicon"
],
[
"microRNA",
null,
null
],
[
"DLK1-DIO3\nRNA cluster",
"mir-127, mir-134, mir-379, mir-409, mir-382, mir-758, mir-\n381, mir-370, mir-654, mir-431",
"includes MEG3 long\nnoncoding RNA; asso-\nciated with MES"
],
[
"\u201cmir-509\u201d",
"mir-509, mir-514, mir-508",
""
],
[
"\u201cmir-144\u201d",
"mir-144, mir-451, mir-486",
"associated with ery-\nthropoiesis"
],
[
"Methylation",
null,
null
],
[
"\u201cRMND1\u201d",
"RMND1-6-151814639, MAP3K7-6-91353911, DNAAF1-\n16-82735714, PTRH2-17-55139429, ZNF143-11-9439170,\ncg03627896 , TAMM41-3-11863582, CDK5-7-150385869,\nOTUB1-11-63510174, AATF-17-32380976",
""
],
[
"M+",
"cg13928306, MTMR11-1-148175405, cg27324619,\nTNKS1BP1-11-56846646, C11orf52-11-111294703,\nIL17RC-3-9934128, cg24765079, ERBB3-12-54759072,\nIL22RA1-1-24342151, C11orf52-11-111294903",
"methylated in infiltrat-\ning lymphocytes"
]
],
"similarity_score": 0.5887573964497042,
"table_image": "images/1306.2584v2/table_0.png",
"page_image": "pages/1306.2584v2/page_5.png"
}
] |