Molecular Nutrition Unit, Department of Food and Nutrition, Technical University of Munich, Germany
Every nutritional process relies on the interplay of a large number of proteins encoded by mRNA molecules that are expressed in a given cell. Alterations of mRNA levels and in turn of the corresponding protein levels (although the two variables do not necessarily change in parallel) are critical parameters in controlling the flux of a nutrient or metabolite through a biochemical pathway. Nutrients and non-nutrient components of foods, diets and lifestyle can affect essentially every step in the flow of genetic information, from gene expression to protein synthesis to protein degradation, thereby altering metabolic functions in the most complex ways. There is no doubt that with the genetic information emerging on a daily basis, we are discovering exciting tools that provide us with insights into the molecular basis of human metabolism under normal as well as pathophysiological conditions. There is also no doubt that the interplay of the rather static mammalian genome with its rapidly changing nutritional environment is one of the most attractive and interesting areas in post-genomic research.
From Gene to Function and from Genomics to Phenomics
Although a huge body of information on the number of mammalian genes, on chromosomal localization of individual genes, their genomic structure and in part also on the functions of the encoded proteins has been gathered, we are far from understanding how these individual factors orchestrate metabolism.
Genomic data contain only limited information about the dynamic behaviour of integrated cellular processes. Nevertheless, recent technological advances have made it possible to analyse the variability and dynamic changes in the genetic response of a cell or organism by determining the expression level of individual RNA molecules or huge sets of mRNA molecules. Whereas genomics describes large-scale DNA sequencing that provides basic genetic information and insights into sequence heterogeneity (i.e. single nucleotide polymorphisms; SNPs) in coding regions of genes as well as in control elements (i.e. promoters), transcriptomics - also called expression profiling - assesses mRNA levels of a few or up to several thousand open reading frames simultaneously in a biological sample; this is done mainly by DNA hybridization arrays and/or by quantitative polymerase chain reaction (PCR) techniques (Celis, 2000; Lockhart and Winzeler, 2000). Proteomics allows the pro-teome - as the protein complement of the genome that is expressed in a cell or an organ - to be identified, and changes in protein expression patterns
©CAB International 2003. Molecular Nutrition (eds J. Zempleni and H. Daniel)
and levels to be determined. Moreover, for individual proteins, post-translational modifications that are crucial for functions, or even amino acid substitutions (polymorphisms) can be detected (Dutt and Lee, 2000; Pandey and Mann, 2000). Applications of these new techniques of genome, transcriptome and proteome analysis are central to the development of nutritional sciences in the next decade and its integration into the rapidly evolving era of functional genomics.
Functional genomics is either based on gene-driven or on phenotype-driven approaches, as shown in Fig. 2.1. The gene-driven approaches use genomic information to identify, clone, express and characterize genes at the molecular level. Pheno-type-driven approaches characterize phenotypes from random mutation screens or naturally occurring variants to identify and clone the gene(s) responsible for the particular phenotype, without knowledge of the underlying molecular mechanisms. Of course, the two strategies are highly complementary at virtually all levels of analysis and lead collectively to the correlation of genotypes and phenotypes. Where nature has not provided inborn errors of metabolism that demonstrate the pheno-typical consequences of individual gene/protein malfunctions, the role of single genes or groups of genes in the make-up of metabolism can be analysed by gene inactivation ('knock-out') or selective expression ('knock-in') and overexpression models employing experimental animals such as fruit flies (Drosophila melanogaster), nematodes (Caenorhabditis elegans), mice, rats and human cell lines. These approaches oftransgenics have already produced a large number of animal lines lacking one or several genes, or overexpressing others. Although very elegant as new genetic tools for understanding metabolism, these manoeuvres unfortunately quite often do not produce an obvious phenotype. This may tell the investigator that the lack of gene function may be compensated for by other mechanisms. More advanced transgenic technologies in animals such as Cre/lox-controlled cell-, organ- and/or time-dependent gene inactivation or induction of expression allow for analysis of phenotypic consequences in even more elegant ways; these techniques appear to be helpful in particular in those situations where simple gene disruption has been lethal for the developing fetus or newborn (Sauer, 1998).
Although functional genomics starts with the classical one-gene approach, it will rapidly move to more system-based 'holistic' levels by employing the high-throughput technologies (DNA micro-arrays and proteomics) in combination with targeted gene deletions or selective overexpression to explore the consequences ofoperational shifts in genetic circuits and cellular systems. This means that we will be facing numerous and huge gene and protein expression databases. To bring meaning and value into these data sets, we require system-based approaches and computational strategies with algorithms that help to describe metabolism (Paton et al, 2000; Tomita, 2001).
Closing the circle: from metabolism to metabolomics
Steady-state levels of human metabolism are determined by the rates of biosynthesis and degradation of proteins (turnover) that function as enzymes, receptors, transporters, channels, hormones and other signalling molecules or that provide structural elements for cells, organs or the skeleton. Between the proteins, there is a variable flow of metabolic intermediates that serve as building blocks for proteins, carbohydrates, lipids or heterooligomers and that provide the fuel for ATP synthesis. Whereas in the past we mainly looked at the phenotypical expression of metabolism by measuring the concentration of a few individual metabolic intermediates, the new molecular tools allow us now to determine every step in the flow of the biological information from DNA to mRNA to proteins and to function. However, biological regulation goes beyond the control of gene expression and protein synthesis: major determinants of metabolism include proteinprotein interactions and alterations of protein activity by metabolic intermediates in terms of kinetic effects and/or allosteric modulation of function. So, whatever information is gathered at the levels of mRNA and protein expression will not allow the prediction of metabolic consequences sufficiently (ter Kuile and Westhoff, 2001). The final stage along the line from gene to mRNA, to protein, to function is therefore the analysis of the pattern and the concentrations of the metabolites that flow between proteins, organelles, cells and organs. Thus, we have to end where we started - looking at metabolites. However, this time, analysis of the entire metabolome
Comparative genomics Identification ot genes with homologous functions
Large mutant collections
Monogenic or complex
Classifications of phenotypes
Analysis of molecular phenofypes
.ï V kr
Modulating gene or protein activity in vivo/ in vitro understanding gene function
Modulating gene or protein activity in vivo/ in vitro understanding gene function
Fig. 2.1. Overview on the two different approaches of functional genomics. One route is guided by the analysis of the phenotype in model organisms or human monogenic diseases to identify the underlying gene and its function. The other route is guided by analysing functions of an individual gene. The main analysis tools and investigative methods are shown with the links between the different levels of information.
as the sum of all detectable low and intermediate molecular weight compounds rather than individual metabolites will be performed by the meta-bolomics approaches. Different physical methods can be used for the comprehensive analysis of metabolites within a biological sample. In most cases, procedures are coupled to classical chro-matographic separation techniques and comprise Fourier transform infrared spectroscopy, electro-spray ionization mass spectrometry (ESI-MS) and nuclear magnetic resonance (NMR) spectroscopy
(Fiehn et al., 2000; Glassbrook et al., 2000). The potential of comprehensive metabolic analysis coupled to statistical methods of cluster analysis for discriminatory phenotype analysis becomes obvious when inspecting research progress achieved in bacterial systems, yeast and plants employing wild-type as well as transgenic organisms (Delneri et al., 2001; Kose et al, 2001). Phenomics (Schilling et al., 1999) finally takes gene function on such a systemic scale. It utilizes all genomic information, all expression information (at the mRNA and protein level) and all metabolites to describe metabolism in the most comprehensive way on the basis of the networks of biological regulation.
Whereas genomics employs the first line classical DNA sequence technology, transcriptome analysis utilizes mainly fluorescence-based detection systems to determine message (mRNA) expression levels in a biological sample. Expression profiling is covered in Chapter 1 and thus will not be addressed here under technological aspects.
The term 'proteome' was introduced as the complement to the genome, and comprises all transcribed and translated open reading frames (protein-encoding regions) in a given genome. Proteome analysis is based on the separation of proteins by two-dimensional polyacrylamide gel electrophoresis (2D-PAGE). Although 2D-PAGE may already appear to be an 'ancient' technique, it provides separation with the highest resolution currently available. However, it is still not easy to attempt to obtain standardized and reproducible separation conditions. Similarly, staining of the proteins in the gel is crucial for the subsequent quality of analysis. Both Coomassie blue and silver staining can be performed, with silver-staining being at least ten times more sensitive. Recently, sensitive fluorescent dyes have been developed that improved the sensitivity further (Patton, 2000). 2D-PAGE separates proteins according to their charge (isoelectric point; pI) by isoelectric focusing (IEF) in the first dimension and according to their size (molecular mass) by SDS-PAGE in the second dimension. It therefore has a unique capacity to resolve complex mixtures of proteins, permitting the simultaneous analysis of hundreds or even thousands of gene products (Gorg et al., 2000). However, not all proteins are resolved and separated equally well by 2D-PAGE. Analysis of very alkaline, hydrophobic and integral membrane proteins as well as high molecular weight proteins is still a problem. In some cases, a pre-fractionation according to cellular compartment (membranes/microsomes, cytosol, mitochondria) or according to protein solubility may be necessary (Blackstock and Weir, 1999; Cordwell et al., 2000). In addition, proteins of low cellular abundance, which may be particularly important in view of their cellular functions (e.g. in signalling pathways), are still very difficult to resolve in the presence of large quantities of housekeeping proteins (Corthals et al., 2000). However, new concepts are constantly being developed and include tagging techniques (Adam et al., 2001) and the enrichment of minor proteins prior to separation in 2D gels.
Two developments have been of central importance in the revival of 2D-PAGE protein separation in proteomics: (i) improvement of highresolution mass spectrometers for ionization and analysis of peptides and proteins based on their molecular mass; and (ii) sequencing of the genomes of various species (including humans) to provide information on coding regions of expressed proteins. Finally, the development of computer algorithms that match uninterpreted (or partially interpreted) mass spectra with the translation of the nucleotide sequence of expressed genes from databases was also important for the advancement of the field of proteomics.
Although the most common procedure for the identification of a protein spot in a gel currently is peptide mapping or 'fingerprint' analysis, there are a number ofother techniques and approaches that can also be applied (Chalmers and Gaskell, 2000; Gevaert and Vandekerckhove, 2000). For example, the proteins may be transferred from the gel on to a membrane to make them accessible to specific antibodies that then detect the proteins by an immunoreaction. Proteins can also be excised from the gel and be submitted to classical Edman degradation with N-terminal sequencing or to ESI-MS analysis to obtain an amino acid sequence; this sequence can then be used to identify the protein by comparison with sequences deposited in databases.
To obtain sequence information, the more rapid and advanced techniques employ peptide mass analysis. Here the protein-containing spots are excised from the gel and the gel piece is altered chemically to make the protein accessible to hydrolysis by a protease such as trypsin. Digestion requires a highly purified trypsin that is not contaminated with enzymes that may cleave peptide bonds containing cationic amino acid residues. Based on the site-specific hydrolysis by trypsin, a characteristic pattern of peptide fragments serves as a peptide mass fingerprint to identify the parent protein.
The mixture of peptides produced by digestion with protease is submitted usually to matrixassisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS) analysis to determine the corresponding peptide masses that are characteristic for a given protein. The mass spectrum obtained is analysed by software that interprets the pattern and predicts the protein by various algorithms, based on a comparison with masses predicted by 'virtual digestion' of identified open reading frames (coding amino acid sequences) in a given genome. Nevertheless, the mass spectra require further careful analysis to ensure the accuracy of the observation. For example, if the measured mass exceeded the predicted mass of a given peptide, one could predict that post-translational modifications of the parent protein may have occurred, such as addition of phosphate groups, hydroxylations at lysine or proline residues, glycosylations or addition of fatty acids. Other deviations of measured from predicted masses may be due to polymorphisms in coding sequences with subtle amino acid substitutions, or even more pronounced with deletions or insertions. Such changes in the primary sequence can be resolved, but success rates depend on the type of substitution, and analysis may require internal peptide Edman sequencing or more advanced mass analysis by ESI-MS (Gaskell, 1997).
Proteome analysis is straightforward if homogenous populations of cultured cells or cell lines are analysed. In contrast, tissue samples contain various cell populations that have different expression profiles that may contribute in various ways to the proteome pattern obtained. Analysis is complicated further by variations in biological samples (i.e. different percentages of individual cell types), which may make analysis a particularly difficult task. Such variations may make it necessary to separate the different cell populations by means of cell-specific surface markers and immunoaffinity techniques or by laser-driven microdissection approaches (Banks etal., 1999; Simone etal., 2000).
The future might hold methods of simplified proteome analysis by using antibody libraries that contain specific antibodies against any expressed open reading frame. This would take proteome analysis to the format of high-throughput microplate assays that permit identification and quantification of essentially every known protein (Brody and Gold, 2000).
The consequences of gene deletions or targeted expression of new genes on the metabolite profiles of organisms are modelled mainly in studies employing bacteria, yeast and plants. Metabolome analysis as a new tool for a comparative display of gene function has the potential not only to provide deeper insights into complex regulatory processes but also to determine phenotypes directly. Automated gas chromatography/mass spectrometry (GC/MS) techniques allow quantification of large numbers of metabolites, the identities of which remain to be determined using chemical procedures. Using these large data sets, various data mining tools for component analysis are employed to assign a given genotype to its characteristic metabolic phenotype (Glassbrook et al., 2000).
Most of the analytical procedures that have been applied to characterize metabolic processes in traditional nutrition studies are based on classical methods of quantitative chemistry. Specific and quantitative analysis is required for focused studies of specific compounds or metabolic pathways. The methods developed have been optimized to produce high-quality data that describe the compounds ofinterest. However, this type ofanalysis is poorly suited to gather information simultaneously on the multitude of metabolites that characterize an organism's nutritional processes.
In metabolic profiling, the concept is different. To monitor hundreds or even thousands of metabolites in parallel, high-throughput techniques are required that allow for screening for relative changes rather than absolute concentrations of compounds. Most analytical techniques for profiling of small molecules consist of a highperformance liquid chromatograph (HPLC) or a gas chromatograph (GC) coupled to a mass spectrometer. Mass spectrometers generally are more sensitive and more selective than any other type of detector. When coupled with the appropriate sample introduction and ionization techniques, mass spectrometers can analyse selectively both organic and inorganic compounds. Nevertheless, prior to detection, the metabolites have to be separated by chromatographic techniques that are coupled on-line to the mass detector. GC is used to separate compounds based on their relative vapour pressures and affinities for the material in the chro-matography column, but is limited to compounds that are volatile and heat stable. Most biological compounds, such as sugars, amino acids and organic acids, are not sufficiently volatile to be separated by GC in their native state and must therefore be derivatized prior to GC separations. HPLC separations are better suited for the analysis of labile and high molecular weight compounds and for the analysis ofnon-volatile polar compounds in their natural form. Although GC- and HPLC-based profiling techniques are not truly quantitative, the compounds detected and their relative amounts may be compared between studies by employing the proper standards. The high-throughput screening with GC-MS and HPLC-MS techniques will also generate large volumes of analytical data that require advanced informatics technologies to organize vast amounts of information.
Metabolite profiling technology will allow information to be gathered on the flow of metabolites through biological pathways and the control of these pathways. In particular, high-resolution 1H-NMR spectroscopy (with the advantage of detection of any proton-containing metabolite) appears to have become increasingly important in metabolite profiling. NMR techniques have been used in the past mainly to analyse metabolite changes in mammalian body fluids and tissues; this method may be extended by detecting other nuclides, e.g. 31P or natural isotopes such as 13C.
When metabolomics is applied to studies where substrates enriched in isotopes such as 13C are administered, metabolite analysis can be taken to a dynamic level by quantification of fluxes (Brenna, 2001). Such automated, biochemical profiling techniques will become an important component of multidisciplinary integrated approaches to metabolic and functional genomics studies.
Current limitations of the technologies when applied to studies in humans
Applications of the technologies of genomics, transcriptomics, proteomics and metabolomics in nutritional studies seem unlimited in terms ofbasic research. In view of applied research interests, these techniques also have a great potential to identify specific markers (biomarkers) that respond to the status of a given nutrient, non-nutrient compound, treatment or diet. Biomarkers of nutrient status may include changes in the levels of individual mRNAs or proteins, but may also include changes in the pattern of a large group of mRNAs or proteins. So far, biomarkers of cellular functions have been identified mainly by rational approaches based on knowledge of metabolism. The new screening approaches are essentially non-logical when analysing several thousands of potentially affected indicator molecules (mRNAs, proteins) simultaneously.
The use of the new technology platforms is essentially unlimited when human cells in culture or model organisms and experimental animals are employed in nutrition research. However, there are limitations with regard to human studies. Expression profiling at the mRNA level is limited by the availability of cells containing sufficient amounts of high-quality RNA for analysis, although the mRNA of a single cell can be used for amplification and quantitation of single genes or sets of genes by PCR techniques. In proteomics, sufficient quantities ofhuman cells for assessing the protein pattern can only be obtained by invasive techniques.
Although certain cells (such as hair follicle cells, skin cells or even exfoliated intestinal cells) may be obtained in sufficient quantities by non-invasive techniques, various types of blood cells are the preferred source of material (RNA and/or proteins) and may be used as 'reporter cells'. They do respond to dietary changes and also have various life expectancies, various gene expression profiles and control systems, and they target various body compartments. In particular, peripheral blood lymphocytes are already used for identification of potential mRNA biomarkers in human studies in response to environmental factors employing DNA arrays (Amundson et al., 2000; Glynne etal., 2000; Marrack etal., 2000). For analysis of SNPs, various sources of DNA can be used; large-scale applications for identification of relevant SNPs are the basis of the rapidly growing field of molecular epidemiology in all areas of biomedical sciences (Haugen, 1999; Tan etal., 1999; Beeley et al., 2000; Perera and Weinstein, 2000; Schork, etal., 2000). Expression profiling, SNP analysis and proteomics are already well established technology platforms in drug discovery studies and are central for pharmacogenomics, a phrase that links genetic heterogeneity (SNPs) and selective expression of drug-metabolizing enzymes to phenotypical differences in the kinetics and dynamics of drug metabolism (Beeley et al., 2000; Rininger et al., 2000; Norton, 2001). To date, some nutrition studies have utilized the new techniques of genomics, proteomics and metabolomics. These studies include application of gene arrays in response to micronutrient status (see Chapter 1) or to caloric restriction (Lee et al., 1999), application of pro-teomics to identify effects of ligands of peroxisomal proliferator receptors in obese mice (Edvardsson et al., 1999) and metabolic phenograms of plasma components and their diet-induced changes in rats (Vigneau-Callahan et al., 2001).
Gene-nutrient interactions are the paradigm for the interplay between the genome and the environment. We are just entering the era of post-genomic research, and there is no doubt that molecular nutritional science is going to be of central interest as nutrients and other food components are the key factors in affecting gene and protein activities. The wealth of genetic information and novel techniques with high-throughput capabilities provide exciting tools for nutrition research. Knowledge of the response of mammalian organisms to changes in their nutritional environment may be gathered at the mRNA and/or protein levels by expression arrays, pro-teome analysis and high-throughput metabolite profiling. As those tools can generate overwhelming data sets, there is clearly an emerging need for bioinformatics in nutritional sciences, but also for nutrition researchers with a good knowledge base in cell biology and biochemistry of metabolism. The ultimate goals in the application of all these techniques are to expand our understanding of metabolism and nutrition and to determine how this relates to health and disease.
Adam, G.C., Cravatt, B.F. and Sorensen, EJ. (2001) Profiling the specific reactivity of the proteome with non-directed activity-based probes. Chemistry and Biology 8, 81-95.
Amundson, S.A., Do, K.T., Shahab, S., Bittner, M., Meltzer, P., Trent, J. and Fornace, A.J. Jr (2000) Identification of potential mRNA biomarkers in peripheral blood lymphocytes for human exposure to ionizing radiation. Radiation Research 154, 342-346.
Banks, R.E., Dunn, MJ., Forbes, M.A., Stanley, A., Pappin, D., Naven, T., Gough, M., Harnden, P. and Selby, P.J. (1999) The potential use of laser capture microdissection to selectively obtain distinct populations ofcells for proteomic analysis -preliminary findings. Electrophoresis 20, 689-700.
Beeley, LJ., Duckworth, D.M. and Southan, C. (2000) The impact ofgenomics on drug discovery. Progress in Medicinal Chemistry 37, 1-43.
Blackstock, W.P. and Weir, M.P. (1999) Proteomics: quantitative and physical mapping of cellular proteins. Trends in Biotechnology 17, 121-127.
Brenna, J.T. (2001) Natural intramolecular isotope measurements in physiology: elements of the case for an effort toward high-precision position-specific isotope analysis. Rapid Communications in Mass Spectrometry 5, 1252-1562.
Brody, E.N. and Gold, L. (2000) Aptamers as therapeutic and diagnostic agents. Journal of Biotechnology 74, 5-13.
Celis, J.E., Kruhoffer, M., Gromova, I., Frederiksen, C., Ostergaard, M., Thykjaer, T., Gromov, P., Yu, J., Palsdottir, H., Magnusson, N. and Orntoft, T.F. (2000) Gene expression profiling: monitoring transcription and translation products using DNA micro-arrays and proteomics. FEBSLetters 480, 2-16.
Chalmers, M.J. and Gaskell, S.J. (2000) Advances in mass spectrometry for proteome analysis. Current Opinion in Biotechnology 11, 384-390.
Cordwell, SJ., Nouwens, A.S., Verrills, N.M., Basseal, D.J. and Walsh, B.J. (2000) Subproteomics based upon protein cellular location and relative solubilities in conjunction with composite two-dimensional electrophoresis gels. Electrophoresis 21, 1094-1103.
Corthals, G.L., Wasinger, V.C., Hochstrasser, D.F. and Sanchez, J.C. (2000) The dynamic range ofprotein expression: a challenge for proteomic research. Electrophoresis 21, 1104-1115.
Delneri, D., Brancia, F.L. and Oliver, S.G. (2001) Towards a truly integrative biology through the functional genomics of yeast. Current Opinion in Biotechnology 12, 87-91.
Dutt, M.J. and Lee, K.H. (2000) Proteomic analysis. Current Opinion in Biotechnology 11, 176-179.
Edvardsson, U., Alexandersson, M., Brockenhuus von Lowenhielm, H., Nystrom, A.C., Ljung, B., Nilsson, F. and Dahllof, B. (1999) A proteome analysis of livers from obese (ob/ob) mice treated with the peroxisome proliferator WY14,643. Electrophoresis 20, 935-942.
Fiehn, O., Kopka, J., Dormann, P., Altmann, T., Trethewey, R.N. and Willmitzer, L. (2000) Metabolite profiling for plant functional genomics. Nature Biotechnology 18, 1157-1161.
Gaskell, S.J. (1997) Electrospray: principles and practice. Journal of Mass Spectrometry 32, 677-688.
Gevaert, K. and Vandekerckhove, J. (2000) Protein identification methods in proteomics. Electrophoresis 21, 1145-1154.
Glassbrook, N., Beecher, C. and Ryals, J. (2000) Metabolic profiling on the right path. Nature Biotechnology 18, 1142-1143.
Glynne, R., Ghandour, G., Rayner, J., Mack, D.H. and Goodnow, C.C. (2000) B-lymphocyte quiescence, tolerance and activation as viewed by global gene expression profiling on microarrays. Immunological Reviews 176, 216-246.
Gorg, A., Obermaier, C., Boguth, G., Harder, A., Scheibe, B., Wildgruber, R. and Weiss, W. (2000) The current state oftwo-dimensional electrophore-sis with immobilized pH gradients. Electrophoresis 21, 1037-1053.
Haugen, A. (1999) Progress and potential of genetic susceptibility to environmental toxicants. Scandinavian Journal of Work Environment and Health 25, 537-540.
Kose, F., Weckwerth, W., Linke, T. and Fiehn, O. (2001) Visualizing plant metabolomic correlation networks using clique-metabolite matrices. Bio-informatics 17, 1198-1208.
Lee, C.K., Klopp, R.G., Weindruch, R. and Prolla, T.A. (1999) Gene expression profile of aging and its retardation by caloric restriction. Science 285, 1390-1393.
Lockhart,DJ. andWinzeler, E.A. (2000) Genomics, gene expression and DNA arrays. Nature 405, 827-836.
Marrack, P., Mitchell, T., Hildeman, D., Kedl, R., Teague, T.K., Bender, J., Rees, W., Schaefer, B.C. and Kappler, J. (2000) Genomic-scale analysis of gene expression in resting and activated T cells. Current Opinion in Immunology 12, 206-209.
Norton, R.M. (2001) Clinical pharmacogenomics: applications in pharmaceutical R&D. DrugDiscovery Today 6, 180-185.
Pandey, A. and Mann, M. (2000) Proteomics to study genes and genomes. Nature 405, 837-846.
Paton, N.W., Khan, S.A., Hayes, A., Moussouni, F., Brass, A., Eilbeck, K., Goble, C.A, Hubbard, SJ. and Oliver, S.G. (2000) Conceptual modelling of genomic information. Bioinformatics 16, 548-557.
Patton, W.F. (2000) Makingblind robots see: the synergy between fluorescent dyes and imaging devices in automated proteomics. Biotechniques 28, 944-948.
Perera, F.P. and Weinstein, I.B. (2000) Molecular epidemiology: recent advances and future directions. Carcinogenesis 21, 517-524.
Rininger,J.A., DiPippo, V.A. and Gould-Rothberg, B.E. (2000) Differential gene expression technologies for identifying surrogate markers of drug efficacy and toxicity. Drug Discovery Today 5, 560-568.
Sauer, B. (1998) Inducible gene targeting in mice using the Cre/lox system. Methods 14, 381-392.
Schilling, C.H, Edwards, J.S. and Palsson, B.O. (1999) Toward metabolic phenomics: analysis ofgenomic data using flux balances. Biotechnology Progress 15, 288-295.
Schork, N.J., Fallin, D. and Lanchbury, J.S. (2000) Single nucleotide polymorphisms and the future ofgenetic epidemiology. Clinical Genetics 58, 250-264.
Simone, N.L., Paweletz, C.P., Charboneau, L., Petricoin, E.F. and Liotta, L.A. (2000) Laser capture microdissection: beyond functional genomics to proteomics. Molecular Diagnosis 5, 301-307.
ter Kuile, B.H. and Westerhoff, H.V. (2001) Transcriptome meets metabolome: hierarchical and metabolic regulation ofthe glycolytic pathway. FEBSLetters 500, 169-171.
Tomita, M. (2001) Whole-cell simulation: a grand challenge of the 21st century. Trends in Biotechnology 19, 205-210.
Vigneau-Callahan, K.E., Shestopalov, A.I., Milbury, P.E., Matson, W.R. and Kristal, B.S. (2001) Characterization ofdiet-dependent metabolic serotypes: analytical and biological variability issues in rats.
Journal of Nutrition 131, 924S-932S.
Was this article helpful?