- Open Access
Identification of frequent cytogenetic aberrations in hepatocellular carcinoma using gene-expression microarray data
© Crawley and Furge, licensee BioMed Central Ltd 2002
- Received: 19 August 2002
- Accepted: 17 October 2002
- Published: 25 November 2002
Hepatocellular carcinoma (HCC) is a leading cause of death worldwide. Frequent cytogenetic abnormalities that occur in HCC suggest that tumor-modifying genes (oncogenes or tumor suppressors) may be driving selection for amplification or deletion of these particular genetic regions. In many cases, however, the gene(s) that drive the selection are unknown. Although techniques such as comparative genomic hybridization (CGH) have traditionally been used to identify cytogenetic aberrations, it might also be possible to identify them indirectly from gene-expression studies. A technique we have called comparative genomic microarray analysis (CGMA) predicts regions of cytogenetic change by searching for regional gene-expression biases. CGMA was applied to HCC gene-expression profiles to identify regions of frequent cytogenetic change and to identify genes whose expression is misregulated within these regions.
Using CGMA, 104 HCC gene-expression microarray profiles were analyzed. CGMA identified 13 regions of frequent cytogenetic change in the HCC samples. Ten of these regions have been detected in previous CGH studies (+lq, -4q, +6p, -8p, +8q, -13q, -16q, -17p, +17q, +20q). CGMA identified three additional regions that have not been previously identified by CGH (+5q, +12q, +19p). Genes located in regions of frequent cytogenetic change were examined for changed expression in the HCC samples.
Our results suggest that CGMA predictions using gene-expression microarray datasets are a practical alternative to CGH profiling. In addition, CGMA might be useful for identifying candidate genes within cytogenetically abnormal regions.
- Comparative Genomic Hybridization
- Pituitary Tumor Transforming Gene
- Comparative Genomic Hybridization Analysis
- Cytogenetic Change
- Comparative Genomic Hybridization Data
Aneuploidy is a common feature of cancer. Genetic alterations such as amplification, deletion, translocation and rearrangement could result in either gain-of-function or loss-of-function mutations in genes that modulate aspects of cell proliferation, differentiation, motility and survival. Whereas cytogenetic profiling techniques, such as comparative genomic hybridization (CGH) , have been useful in finding genetic abnormalities, other experimental approaches are frequently used to identify which specific gene(s) drive selection for the genetic aberration and contribute most to tumor progression. Common gene identification techniques include determining if a candidate gene contains a sequence mutation and/or determining if the candidate gene or gene product is abnormally expressed. As mutation analysis and protein expression studies are time-consuming, increasingly high-throughput gene-expression profiling is being used to identify abnormally expressed genes within a region of cytogenetic change [2,3,4,5,6].
Recently, several groups have observed that chromosomal changes can lead to regional biases in gene-expression values both in yeast (Saccharomyces cerevisiae) and in human tumors and tumor-derived cell lines [2,3,7,8]. These studies suggest that a fraction of gene-expression values (15-25%) are regulated in concordance with gene dosage. A computational technique termed comparative genomic microarray analysis (CGMA) has previously been used to identify regions of allelic imbalance indirectly from gene-expression profiles of human tumors . CGMA predicts chromosomal amplifications and deletions by organizing gene-expression data by genomic mapping location and scanning for regions that contain a statistically significant number of gene-expression values that change in the same relative direction. In this study, we apply CGMA analysis to a large hepatocellular carcinoma microarray dataset to demonstrate its validity as an alternative to CGH and to identify candidate genes in regions of frequent cytogenetic change.
Primary liver cancer in adults is the sixth most common form of cancer and the fourth leading cause of death from cancer worldwide [9,10]. Through the examination of hepatitis B virus (HBV) - and hepatitis C virus (HCV)-induced tumors, two landmark CGH studies have suggested that a subset of cytogenetic changes frequently occurs in HCC [11,12]. These include frequent gain of chromosomes 1q, 6p, 8q, 17q and 20q and frequent loss of chromosomes 1p, 4q, 6q, 8p, 13, 16 and 17p [11,12]. In particular, gain of chromosomes 1q and 8q has been associated with the early development of HCC , whereas loss of chromosome 4q has been linked to increased aggressiveness of established tumors . To determine whether gene-expression data could be used to identify cytogenetic changes accurately, we applied CGMA to a microarray dataset of HCC tumors and compared the CGMA predictions to existing CGH data. For HCC, CGMA was able to predict nearly all chromosomal aberrations identified previously by CGH. In addition, from the gene-expression data we also identified a set of genes whose expression values change most within the regions of cytogenetic change. These genes may represent candidate genes whose expression changes drive selection for chromosomal gains or losses.
CGMA predictions of cytogenetic changes
Comparison to previous HCC studies
CGMA predictions of multifocal tumors
Identification of candidate genes in regions of frequent cytogenetic change
Identification of HCC candidate genes located within a region of chromosome 8q
Unknown (protein for mgc:14128)
cDNA flj14825 fis, clone ovarc1000781
Zinc finger homeobox protein zhx1 (zhx1 protein)
Similar to riken cDNA 2610509g12 gene
Hypothetical 23.7 kDa protein
Similar to Homo sapiens F-box protein fbx25 (fbx25)
Annexin a13 (annexin XIII)
cDNA FLJ20772 FIS, Clone COL06053
NADH-ubiquinone oxidoreductase b22 subunit
Hypothetical protein kiaa0429
cDNA flj32440 fis, clone skmus2001492
G-protein-coupled receptor induced protein Gig2
Contains a reverse transcriptase domain
Estradiol 17 beta dehydrogenase 4 EC 1.1.1
Hypothetical 23.7 kDa protein
Myc proto-oncogene protein (c-Myc)
Contains a reverse transcriptase domain
Misregulated genes located in regions of frequent cytogenetic aberrations
Hypothetical protein NUF2R
Ubiquitin-conjugating enzyme E2C
Chromosome 20 open reading frame 1
Pituitary tumor-transforming 1
HSPC150 protein similar to ubiquitin-conjugating enzyme
Hypothetical protein MGC5576
Adenylyl cyclase-associated protein 2
Ribonucleotide reductase M2 polypeptide
Regulator of G-protein signalling 5
Glutathione S-transferase A4
Chromosome 1 open reading frame 2
Alcohol dehydrogenase 4 (class II), pi polypeptide
RNA helicase-related protein
deltex (Drosophila) homolog 1
Early growth response 1
Extracellular matrix protein 1
Lymphocyte antigen 6 complex, locus E
KIAA0711 gene product
Sex hormone-binding globulin
Pleckstrin homology-like domain, family A, member 1
Baculoviral IAP repeat-containing 5 (survivin)
Small inducible cytokine A2
Platelet-derived growth factor receptor, alpha polypeptide
Non-metastatic cells 3, protein expressed in
Cyclin-dependent kinase 5, regulatory subunit 1 (p35)
CGMA prediction software
To assist in identifying regions of unidirectional gene-expression bias, we have constructed a web-based program that processes two-color gene-expression data and identifies genomic regions that contain gene-expression biases. The input for this program is a simple tab-delimited gene-expression matrix file consisting of columns for the probe sequence identifier, probe name, and gene-expression ratios. Because different microarray technologies use different identifiers to describe the microarray probe, the program translates probe sequence identifiers (ids) such as GenBank accession numbers and UniGene cluster ids to Ensembl transcript ids using precompiled sequence comparisons. After data analysis, a summary table is displayed showing chromosomal regions that show significant (α ≤ 0.05) unidirectional gene-expression bias highlighted in either red or green, indicating either increased or decreased expression biases, respectively. The program can also send several output files to the user via e-mail. These files include a summary report that contains the z-statistic for each chromosomal region (positive for upregulated regions and negative for downregulated regions) and a list of genes located in regions of frequent cytogenetic change. The program is available at .
In this study we have used gene-expression profiling data to predict cytogenetic changes that frequently occur in HCC. Two landmark CGH analyses identified 12 different regions of frequent imbalance. However, one study found 8 regions and the other study found 11 [11,12]. Five of these 12 regions were not found in both experiments. CGMA successfully identified 10 of 12 regions previously distinguished by CGH. CGMA also detected three regions that have not been implicated by these CGH studies. On average however, 22% of genomic regions indentified in a particular HCC CGH study are not constantly identified in other studies. Therefore, the three inconsistent CGMA predictions (3 of 13; 23%) are comparable to the inconsistencies between independent CGH studies for HCC.
Three additional regions were identified by CGMA that were not identified by CGH. While these CGMA-predicted regions were near the 35% cutoff for detection, they could represent other regions of allelic imbalance yet to be detected by CGH. It is also possible that biological mechanisms other than cytogenetic change could influence expression in large genomic regions and produce regional gene-expression biases. Additional molecular genetic work will be required to resolve these differences.
If CGH data are not available for a particular cancer type, but gene-expression profiling data are, then CGMA could allow rapid prediction of the cytogenetic abnormalities that frequently occur within that cancer type. Moreover, in instances where gene-expression profiling reveals previously unrecognized cancer subtypes, CGMA could determine whether cytogenetic differences are responsible for these different subtypes. In cancer types where traditional cytogenetic profiling has already been carried out, CGMA predictions could serve to confirm existing cytogenetic profiling data and be used further to examine candidate genes whose expression changes most within a region of frequent cytogenetic change. In this way CGMA can be combined with the candidate gene approach to identify genes that are directly involved in tumor progression.
CGMA can be used to indicate chromosomal imbalances by detecting chromosomal regions that contain a disproportionate number of gene-expression values that change in the same relative direction. This analysis provides good evidence that CGMA is a practical alternative to CGH cytogenetic profiling when gene-expression profiling data is available.
Normalization and filtering
Normalized, log-transformed gene-expression data for 104 HCC samples and 76 corresponding non-cancerous liver gene-expression profiles  were obtained from the Stanford Microarray Database . Genes that were present in at least 75% of samples (10,037 genes) were used for further analysis. In this study, both the tumor samples and normal tissue samples were compared to a pooled cell-line reference . To allow comparison of tumor gene-expression values to gene-expression values from surrounding non-cancerous tissue, new gene expression ratios, tumor verse normal (T/N), were estimated. To create the new ratios, log-transformed non-cancerous tissue ratios (N/U) were subtracted from the log-transformed HCC tissue ratios (T/U) for each gene such that log2(T/N) = log2(T/U) - log2(N/U). If an HCC sample did not have a corresponding non-cancerous sample, the global mean of the non-cancerous tissue gene-expression ratios were used.
To identify regional gene-expression biases, gene-expression values that map within a given chromosomal arm were collected and a sign test for a one-sample mean/median was used to determine whether a significant upward or downward bias was present in the expression values. An exception was made for chromosomes 13-16, 21 and 22. These chromosomes are more telocentric and therefore only their q-arms were tested for expression biases. Sequence comparisons were used to map microarray probe sequences (the sequences that are placed on the microarray) to predicted Ensembl transcripts (Ensembl version 6.28) . Included in the Ensembl transcript annotations are chromosomal mapping locations at base-pair resolution. Redundancy introduced by replicate probes on the array and/or multiple probes mapping to the same gene were eliminated by averaging expression values that map to identical transcripts. Of the filtered set of 10,037 genes, 6,274 genes (63%) were unique and had associated genomic mapping information.
A sign test for the one-sample mean/median was used to determine whether a significant number of genes that map to a given chromosomal region change in a unidirectional manner. The algorithm scores a gene as up (+) or down (-) regulated if the magnitude of the expression value change is at least 1.8-fold. The sign test computes the probability, in the form of a z-statistic, of finding x upregulated genes out of n genes that change in a given genomic region. For simplicity, the z-statistic is computed using the normal approximation to the binomial distribution such that z = (2x - n)/sqrt(n). Genomic regions that contained less than 15 changed gene-expression values were excluded from further analysis. On average, 160 gene-expression values were located to each genomic region. The sign test z-statistic can be converted to a significance value (α) based on the two-tailed z-statistic (zα/2) critical values. For example, if z = 1.96, then α = 0.05; if z = 2.58 then α = 0.01, and so on.
Funding was through the generosity of the Van Andel Research Institute and Michigan Center for Biological Information (MCBI). We specially thank Ramsi Haddad for helpful discussion in the preparation of this manuscript.
- Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F, Pinkel D: Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science. 1992, 258: 818-821.PubMedView ArticleGoogle Scholar
- Virtaneva K, Wright FA, Tanner SM, Yuan B, Lemon WJ, Caligiuri MA, Bloomfield CD, de La Chapelle A, Krahe R: Expression profiling reveals fundamental biological differences in acute myeloid leukemia with isolated trisomy 8 and normal cytogenetics. Proc Natl Acad Sci USA. 2001, 98: 1124-1129. 10.1073/pnas.98.3.1124.PubMedPubMed CentralView ArticleGoogle Scholar
- Phillips JL, Hayward SW, Wang Y, Vasselli J, Pavlovich C, Padilla-Nash H, Pezullo JR, Ghadimi BM, Grossfeld GD, Rivera A, et al: The consequences of chromosomal aneuploidy on gene expression profiles in a cell line model for prostate carcinogenesis. Cancer Res. 2001, 61: 8143-8149.PubMedGoogle Scholar
- Monni O, Barlund M, Mousses S, Kononen J, Sauter G, Heiskanen M, Paavola P, Avela K, Chen Y, Bittner ML, Kallioniemi A: Comprehensive copy number and gene expression profiling of the 17q23 amplicon in human breast cancer. Proc Natl Acad Sci USA. 2001, 98: 5711-5716. 10.1073/pnas.091582298.PubMedPubMed CentralView ArticleGoogle Scholar
- Platzer P, Upender MB, Wilson K, Willis J, Lutterbaugh J, Nosrati A, Willson JK, Mack D, Ried T, Markowitz S: Silence of chromosomal amplifications in colon cancer. Cancer Res. 2002, 62: 1134-1138.PubMedGoogle Scholar
- Xu XR, Huang J, Xu ZG, Qian BZ, Zhu ZD, Yan Q, Cai T, Zhang X, Xiao HS, Qu J, et al: Insight into hepatocellular carcinogenesis at transcriptome level by comparing gene expression profiles of hepatocellular carcinoma with those of corresponding non-cancerous liver. Proc Natl Acad Sci USA. 2001, 98: 15089-15094. 10.1073/pnas.241522398.PubMedPubMed CentralView ArticleGoogle Scholar
- Hughes TR, Roberts CJ, Dai H, Jones AR, Meyer MR, Slade D, Burchard J, Dow S, Ward TR, Kidd MJ, et al: Widespread aneuploidy revealed by DNA microarray expression profiling. Nat Genet. 2000, 25: 333-337. 10.1038/77116.PubMedView ArticleGoogle Scholar
- Haddad R, Furge KA, Miller J, Schoumans J, Haab B, Teh B, Barr L, Webb C: Genomic profiling and cDNA microarray analysis of human colon adenocarcinoma and associated peritoneal metastasis reveals consistent cytogenetic and transcriptional aberrations associated with progression of multiple metastases. Appl Genomics Proteomics. 2002, 1: 123-134.Google Scholar
- Pisani P, Parkin DM, Ferlay J: Estimates of the worldwide mortality from eighteen major cancers in 1985. Implications for prevention and projections of future burden. Int J Cancer. 1993, 55: 891-903.PubMedView ArticleGoogle Scholar
- Parkin DM, Pisani P, Ferlay J: Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer. 1993, 54: 594-606.PubMedView ArticleGoogle Scholar
- Marchio A, Meddeb M, Pineau P, Danglot G, Tiollais P, Bernheim A, Dejean A: Recurrent chromosomal abnormalities in hepatocellular carcinoma detected by comparative genomic hybridization. Genes Chromosomes Cancer. 1997, 18: 59-65. 10.1002/(SICI)1098-2264(199701)18:1<59::AID-GCC7>3.0.CO;2-0.PubMedView ArticleGoogle Scholar
- Wong N, Lai P, Lee SW, Fan S, Pang E, Liew CT, Sheng Z, Lau JW, Johnson PJ: Assessment of genetic changes in hepatocellular carcinoma by comparative genomic hybridization analysis: relationship to disease stage, tumor size, and cirrhosis. Am J Pathol. 1999, 154: 37-43.PubMedPubMed CentralView ArticleGoogle Scholar
- Chen X, Cheung ST, So S, Fan ST, Barry C, Higgins J, Lai K, Dudoit S, Ng I, Rijn M, et al: Gene expression patterns in human liver cancer. Mol Biol Cell. 2002, 13: 1929-1939. 10.1091/mbc.02-02-0023..PubMedPubMed CentralView ArticleGoogle Scholar
- Sherlock G, Hernandez-Boussard T, Kasarskis A, Binkley G, Matese JC, Dwight SS, Kaloper M, Weng S, Jin H, Ball CA, et al: The Stanford Microarray Database. Nucleic Acids Res. 2001, 29: 152-155. 10.1093/nar/29.1.152.PubMedPubMed CentralView ArticleGoogle Scholar
- Wong N, Lai P, Pang E, Fung LF, Sheng Z, Wong V, Wang W, Hayashi Y, Perlman E, Yuna S, et al: Genomic aberrations in human hepatocellular carcinomas of differing etiologies. Clin Cancer Res. 2000, 6: 4000-4009.PubMedGoogle Scholar
- Kusano N, Shiraishi K, Kubo K, Oga A, Okita K, Sasaki K: Genetic aberrations detected by comparative genomic hybridization in hepatocellular carcinomas: their relationship to clinicopathological features. Hepatology. 1999, 29: 1858-18562.PubMedView ArticleGoogle Scholar
- Marchio A, Pineau P, Meddeb M, Terris B, Tiollais P, Bernheim A, Dejean A: Distinct chromosomal abnormality pattern in primary liver cancer of non-B, non-C patients. Oncogene. 2000, 19: 3733-3738. 10.1038/sj.onc.1203713.PubMedView ArticleGoogle Scholar
- Guan XY, Fang Y, Sham JS, Kwong DL, Zhang Y, Liang Q, Li H, Zhou H, Trent JM: Recurrent chromosome alterations in hepatocellular carcinoma detected by comparative genomic hybridization. Genes Chromosomes Cancer. 2000, 29: 110-116. 10.1002/1098-2264(2000)9999:9999<::AID-GCC1022>3.0.CO;2-V.PubMedView ArticleGoogle Scholar
- Zondervan PE, Wink J, Alers JC, Ijzermans JN, Schalm SW, de Man RA, van Dekken H: Molecular cytogenetic evaluation of virus-associated and non-viral hepatocellular carcinoma: analysis of 26 carcinomas and 12 concurrent dysplasias. J Pathol. 2000, 192: 207-215. 10.1002/1096-9896(2000)9999:9999<::AID-PATH690>3.0.CO;2-#.PubMedView ArticleGoogle Scholar
- Tornillo L, Carafa V, Richter J, Sauter G, Moch H, Minola E, Gambacorta M, Bianchi L, Vecchione R, Terracciano LM: Marked genetic similarities between hepatitis B virus-positive and hepatitis C virus-positive hepatocellular carcinomas. J Pathol. 2000, 192: 307-312. 10.1002/1096-9896(2000)9999:9999<::AID-PATH706>3.0.CO;2-O.PubMedView ArticleGoogle Scholar
- Collonge-Rame MA, Bresson-Hadni S, Koch S, Carbillet JP, Blagosklonova O, Mantion G, Miguet JP, Heyd B, Bresson JL: Pattern of chromosomal imbalances in non-B virus related hepatocellular carcinoma detected by comparative genomic hybridization. Cancer Genet Cytogenet. 2001, 127: 49-52. 10.1016/S0165-4608(00)00421-0.PubMedView ArticleGoogle Scholar
- Niketeghad F, Decker HJ, Caselmann WH, Lund P, Geissler F, Dienes HP, Schirmacher P: Frequent genomic imbalances suggest commonly altered tumour genes in human hepatocarcinogenesis. Br J Cancer. 2001, 85: 697-704. 10.1054/bjoc.2001.1963.PubMedPubMed CentralView ArticleGoogle Scholar
- Koo SH, Ihm CH, Kwon KC, Park JW, Kim JM, Kong G: Genetic alterations in hepatocellular carcinoma and intrahepatic cholangiocarcinoma. Cancer Genet Cytogenet. 2001, 130: 22-28. 10.1016/S0165-4608(01)00460-5.PubMedView ArticleGoogle Scholar
- Kitay-Cohen Y, Amiel A, Ashur Y, Fejgin MD, Herishanu Y, Afanasyev F, Bomstein Y, Lishner M: Analysis of chromosomal aberrations in large hepatocellular carcinomas by comparative genomic hybridization. Cancer Genet Cytogenet. 2001, 131: 60-64. 10.1016/S0165-4608(01)00492-7.PubMedView ArticleGoogle Scholar
- Chang J, Kim NG, Piao Z, Park C, Park KS, Paik YK, Lee WJ, Kim BR, Kim H: Assessment of chromosomal losses and gains in hepatocellular carcinoma. Cancer Lett. 2002, 182: 193-202. 10.1016/S0304-3835(02)00083-6.PubMedView ArticleGoogle Scholar
- Balsara BR, Pei J, De Rienzo A, Simon D, Tosolini A, Lu YY, Shen FM, Fan X, Lin WY, Buetow KH, et al: Human hepatocellular carcinoma is characterized by a highly consistent pattern of genomic imbalances, including frequent loss of 16q23.1-24.1. Genes Chromosomes Cancer. 2001, 30: 245-253. 10.1002/1098-2264(2000)9999:9999<::AID-GCC1083>3.0.CO;2-M.PubMedView ArticleGoogle Scholar
- Zimonjic DB, Keck CL, Thorgeirsson SS, Popescu NC: Novel recurrent genetic imbalances in human hepatocellular carcinoma cell lines identified by comparative genomic hybridization. Hepatology. 1999, 29: 1208-1214.PubMedView ArticleGoogle Scholar
- Harada T, Shiraishi K, Kusano N, Umayahara K, Kondoh S, Okita K, Sasaki K: Evaluation of the reliability of chromosomal imbalances detected by combined use of universal DNA amplification and comparative genomic hybridization. Jpn J Cancer Res. 2000, 91: 1119-1125.PubMedView ArticleGoogle Scholar
- Rao UN, Gollin SM, Beaves S, Cieply K, Nalesnik M, Michalopoulos GK: Comparative genomic hybridization of hepatocellular carcinoma: correlation with fluorescence in situ hybridization in paraffin-embedded tissue. Mol Diagn. 2001, 6: 27-37. 10.1054/modi.2001.22021.PubMedView ArticleGoogle Scholar
- Takeo S, Arai H, Kusano N, Harada T, Furuya T, Kawauchi S, Oga A, Hirano T, Yoshida T, Okita K, Sasaki K: Examination of oncogene amplification by genomic DNA microarray in hepatocellular carcinomas: comparison with comparative genomic hybridization analysis. Cancer Genet Cytogenet. 2001, 130: 127-132. 10.1016/S0165-4608(01)00479-4.PubMedView ArticleGoogle Scholar
- Kusano N, Okita K, Shirahashi H, Harada T, Shiraishi K, Oga A, Kawauchi S, Furuya T, Sasaki K: Chromosomal imbalances detected by comparative genomic hybridization are associated with outcome of patients with hepatocellular carcinoma. Cancer. 2002, 94: 746-751. 10.1002/cncr.10254.PubMedView ArticleGoogle Scholar
- Yasui K, Arii S, Zhao C, Imoto I, Ueda M, Nagai H, Emi M, Inazawa J: TFDP1, CUL4A, and CDC16 identified as targets for amplification at 13q34 in hepatocellular carcinomas. Hepatology. 2002, 35: 1476-1484.PubMedView ArticleGoogle Scholar
- Chen X, Cheung ST, So S, Fan ST, Barry C, Higgins J, Lai KM, Ji J, Dudoit S, Ng IO, et al: Gene expression patterns in human liver cancers. Mol Biol Cell. 2002, 13: 1929-1939. 10.1091/mbc.02-02-0023..PubMedPubMed CentralView ArticleGoogle Scholar
- Wong N, Lam WC, Lai PB, Pang E, Lau WY, Johnson PJ: Hypomethylation of chromosome 1 heterochromatin DNA correlates with q-arm copy gain in human hepatocellular carcinoma. Am J Pathol. 2001, 159: 465-471.PubMedPubMed CentralView ArticleGoogle Scholar
- Fujiwara Y, Ohata H, Kuroki T, Koyama K, Tsuchiya E, Monden M, Nakamura Y: Isolation of a candidate tumor suppressor gene on chromosome 8p21.3-p22 that is homologous to an extracellular domain of the PDGF receptor beta gene. Oncogene. 1995, 10: 891-895.PubMedGoogle Scholar
- Kuroki T, Fujiwara Y, Nakamori S, Imaoka S, Kanematsu T, Nakamura Y: Evidence for the presence of two tumour-suppressor genes for hepatocellular carcinoma on chromosome 13q. Br J Cancer. 1995, 72: 383-385.PubMedPubMed CentralView ArticleGoogle Scholar
- CGMA - comparative analysis of microarray data. [http://www.vai.org/vari/downloads/development/clam-0.2.pl]
- Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998, 95: 14863-14868. 10.1073/pnas.95.25.14863.PubMedPubMed CentralView ArticleGoogle Scholar