REVIEWS

Microarrays: biotechnologys discovery

platform for functional genomics
Mark Schena, Renu A. Heller, Thomas P. Theriault, Ken Konrad, Eric Lachenmeier and Ronald W. Davis

Advances in microarray technology enable massive parallel mining of biological data, with biological chips providing
hybridization-based expression monitoring, polymorphism detection and genotyping on a genomic scale. Microarrays
containing sequences representative of all human genes may soon permit the expression analysis of the entire human genome
in a single reaction. These genome chips will provide unprecedented access to key areas of human health, including disease
prognosis and diagnosis, drug discovery, toxicology, aging, and mental illness. Microarray technology is rapidly becoming a
central platform for functional genomics.

B
iological research can be viewed as an information set of tools that allow researchers to link hypothesis
science. Understanding the methodological testing and data (Fig. 1a). In the appropriate method-
architecture on which the discipline operates ological context, data from chip-based experiments can
facilitates the efficient gathering of biological infor- provide significant quantitative information about
mation (data). This hierarchy can be arranged in the important cellular pathways and processes. Microarrays
form of an epistemological pyramid, with the tiers allow the accumulation of large amounts of functional-
demarcated according to their degree of conceptual genomic information by enabling the global ordering
abstraction (Fig. 1a). In biological research, hypotheses of molecules in a parallel fashion (Fig. 1c). Proper
are formulated using the most abstract components of experimental design allows the delineation of respec-
the pyramid, such as world view and theory, and then tive internal and external contributions to the physio-
tested using a large body of rules that govern the selec- logical state (Fig. 1c).
tion and use of methods and tools best suited for All microarray assays contain five discrete experi-
addressing a particular biological question (Fig. 1a). mental steps biological query, sample preparation,
Hypothesis testing ultimately results in the accumu- biochemical reaction, detection and data visualization
lation of data, which, in turn, reshape the world view and modelling1. Under appropriate experimental con-
and theory, leading to subsequent rounds of hypothesis ditions1, chips can provide a quantitative measure of the
formulation and testing, and an enhanced understanding molecules present in biochemical extracts. Microarrays
of a biological question. of complementary DNA (cDNA) sequences, for
One aspect of world view focuses on the notion that example, allow hybridization-based expression moni-
an ecosystem contains interactive components that can toring of the cognate genes26. In these assays, steady-
be arranged according to their biochemical complex- state mRNA levels are deduced from the fluorescence
ity (Fig. 1b). The biochemical-complexity hierarchy intensity at each position on the microarray26.
helps to underscore the fact that biological systems are
highly interactive. Complex organisms such as humans, Advantages of chip assays
for example, exist as a culture that is part of a much Microarray assays are rooted in early biochemical
larger ecosystem; importantly, phenotype is deter- experiments on solid surfaces79. Although reminiscent
mined by a combination of genetic and environmen- of filter-based assays10, chip assays are a fundamental
tal factors (Fig. 1b). Understanding human behaviour, departure from techniques that employ porous mem-
disease and other health issues thus requires more than branes. Chips allow true parallelism, miniaturization,
just a knowledge of genes and genomes: one must multiplexing and automation, and these key features
understand the cellular, physiological, cultural and eco- provide a set of performance specifications26 that
logical context in which genomic instructions are being cannot be achieved with the earlier technologies.
read. Microarray assays allow massive parallel data acquisi-
The rate at which biological information is acquired tion and analysis. Parallelism greatly increases the speed
depends, in part, on the research tools employed. Other of experimental progress and allows meaningful com-
factors being equal, better tools yield better data. parisons to be made between the genes or gene prod-
Microarray technology represents a powerful new ucts represented in the microarray. Microarray assays
may eventually allow the analysis of the entire human
genome in a single reaction, and recent gene-expression
M. Schena and R. W. Davis are at the Department of Biochemistry,
Beckman Center, Stanford University Medical Center, Stanford, CA
experiments in yeast represent an important step
94305, USA. R. A. Heller is at Roche Bioscience, 3401 Hillview towards this goal1113.
Avenue, Palo Alto, CA 94304, USA. T. P. Theriault, K. Konrad Miniaturization of conventional assays is a general
and E. Lachenmeier are at Incyte Pharmaceuticals, 3174 Porter Drive, trend in biomedical research14. Microscale assays reduce
Palo Alto, CA 94304, USA. reagent consumption, minimize reaction volumes,

TIBTECH JULY 1998 (VOL 16) Copyright 1998, Elsevier Science Ltd. All rights reserved. 0167 7799/98/$19.00. PII: S0167-7799(98)01219-0 301
REVIEWS

the accuracy of comparative analysis by eliminating

a World view complicating factors such as chip-to-chip variation,
discrepancies in reaction conditions and other short-
Theory comings inherent in comparing separate experiments.
Rules The multiplexed format has already found uses in
expression analysis26,1113, genotyping16 and DNA
Methods resequencing17.
Microarrays
Tools Advanced manufacturing technologies permit the
mass production of biological chips (biochips), and
Data automation is increasing the proliferation of microarray
assays by ensuring their quality, availability and afford-
b ability; as a result, biochips may eventually become
Extrinsic forces commodity items akin to microchips in the computer
industry. Because of the predicted central role of
Ecosystem microarrays in biomedical research, some experts
Culture believe that the biochip revenues will eventually eclipse
the sales of computer chips.
Human beings Organism Human beings
Organ Microarray technology
Microarray-manufacturing technologies fall into two
Genome main categories synthesis and delivery. In the syn-
thesis approaches, microarrays are prepared in a step-
Gene
wise fashion by the in situ synthesis of nucleic acids and
nt other biopolymers from biochemical building blocks.
Intrinsic forces With each round of synthesis, nucleotides are added to
growing chains until the desired length is achieved. The
c delivery technologies, by contrast, use the exogenous
Intrinsic Extrinsic deposition of preprepared biochemical substances for
forces forces chip fabrication. Molecules such as cDNAs are ampli-
fied by PCR and purified, and small quantities are
deposited onto known locations using a variety of
Biological delivery technologies.
system The key technical parameters for evaluating both the
Microarray
synthesis and delivery technologies include microarray
density and design, biochemical composition and ver-
satility, reproducibility, throughput, quality, cost and
Figure 1 ease of prototyping. Three types of advanced tech-
Methodological principles of microarray assays. (a) Methodological nologies have emerged as early favourites in automated
pyramid for microarray analysis. Levels in the pyramid are arranged microarray production (Fig. 2).
in descending order of abstraction (solid arrow). Tools such as
microarrays, scanners, software and biological kits occupy the sec- Photolithography
ond tier (open arrow). (b) Biological-information hierarchy. Levels in One novel synthesis technology, developed by Fodor
the pyramid are arranged in descending order of biochemical com- and colleagues (Affymetrix, Santa Clara, CA, USA)
plexity. Intact organisms, such as humans, occupy the fifth tier of combines photolithography technology from the semi-
the pyramid. Tiers above (ecosystem and culture) and below (organ, conductor industry with DNA-synthetic chem-
genome, gene, nucleotide [nt]) this tier produce extrinsic (solid istry4,7,13,14,17 to enable high-density oligonucleotide-
arrow) and intrinsic (open arrow) forces acting on organisms, respec- microarray manufacture (Fig. 2a). A key advantage of
tively. (c) Microarrays in systems analysis. Intrinsic and extrinsic this approach over nonsynthetic methods is that
forces affect biological processes such as gene expression, which photoprotected versions of the four DNA building
can be examined with microarrays. blocks allow chips to be manufactured directly from
sequence databases4,7,13,14,17, thereby removing the
increase the sample concentration and accelerate the uncertain and burdensome aspects of sample handling
reaction kinetics. Chip-based assays allow sensitive and and tracking. Another advantage of the photolitho-
rapid data detection with either confocal scan- graphic approach is that the use of synthetic reagents
ners27,1113 or cameras equipped with charged-coupled minimizes chip-to-chip variation by ensuring a high
devices1. Although current microarray assays focus on degree of precision in each coupling cycle. One disad-
nucleic acid hybridization, future studies will undoubt- vantage of this approach is, however, the need for pho-
edly involve the parallel analysis of proteins, lipids, tomasks, which are expensive and time-consuming to
carbohydrates and small molecules. design and build.
Multiplexing, the process by which multiple samples Affymetrix chips currently contain as many as
are analysed in a single assay, is another enabling feature 400 000 groups of oligonucleotides or features in an
of the microarray approach. Novel labelling and detec- area of ~1.6 cm2, with each feature containing approx-
tion methods, such those involving multicolour fluo- imately ten million oligonucleotides of a given
rescence15, allow comparisons of multiple samples sequence. Oligonucleotides are anchored at the 3 end,
to be made on a single chip. Multiplexing increases thereby maximizing the availability of single-stranded

302 TIBTECH JULY 1998 (VOL 16)

 REVIEWS

a M2
Shine X X
M1 M1
light Couple Repeat
X X X X X X X X A A

Microarray
b

Touch Move
surface pins Repeat

Microarray
c

Deliver Move
drop jets Repeat

Microarray

Figure 2
Microarray technologies; three approaches to microarray manufacturing are depicted. (a) Photolithography: a glass wafer modified with
photolabile protecting groups (X) is selectively activated for DNA synthesis by shining light through a photomask (M1). The wafer is then
flooded with a photoprotected DNA base (AX), resulting in spatially defined coupling on the chip surface. A second photomask (M2) is used
to deprotect defined regions of the wafer. Repeated deprotection and coupling cycles enable the preparation of high-density oligonucleotide
microarrays. (b) Mechanical microspotting: a biochemical sample is loaded into a spotting pin by capillary action, and a small volume is
transferred to a solid surface by physical contact between the pin and the solid substrate. After the first spotting cycle, the pin is washed
and a second sample is loaded and deposited to an adjacent address. Robotic control systems and multiplexed printheads allow automated
microarray fabrication. (c) Ink jetting: a biochemical sample is loaded into a miniature nozzle equipped with a piezoelectric fitting (rectan-
gles) and an electrical current is used to expel a precise amount of liquid from the jet onto the substrate. After the first jetting step, the jet
is washed and a second sample is loaded and deposited to an adjacent address. A repeated series of cycles with multiple jets enables rapid
microarray production.

nucleic acid for hybridization. Future modifications of premade biochemical substances onto solid surfaces
the Affymetrix approach, such as the use of acid-resist (Fig. 2b). Printing is accomplished by direct surface
technology, may allow microarray manufacturing with- contact between the printing substrate and a delivery
out the need for photomasks (J. Beecher, pers. com- mechanism that contains an array of tweezers, pins or
mun.). Steady improvements in coupling efficiency, capillaries that serve to transfer the biochemical sam-
density and biochemical diversity will ensure the via- ples to the surface (Fig. 2b).
bility of the Affymetrix platform in a competitive Some of the advantages of the microspotting tech-
micorarray marketplace. nologies include ease of prototyping and therefore rapid
implementation, low cost and versatility. One disad-
Mechanical microspotting vantage of microspotting is that each sample must be
A second robust set of technologies are the mechan- synthesized, purified and stored prior to microarray
ical microspotting approaches, an original version of fabrication. The microspotted microarrays currently
which was developed by Shalon and Brown2,16 and later manufactured at Synteni contain as many as 10 000
commercialized at Synteni (Fremont, CA, USA). groups of cDNAs in an area of ~3.6 cm2; each cDNA
Microspotting, a miniaturized version of earlier DNA- feature permits the expression monitoring of its cog-
spotting techniques10, encompasses a family of related nate human gene. A set of four Synteni microarrays
deposition technologies that enable automated should thus allow the expression monitoring of
microarray production by printing small quantities of ~40 000 human genes, the number of unique expressed

TIBTECH JULY 1998 (VOL 16) 303

REVIEWS

Table 1. Microarray Industry

Company Contact information Key products and services

Affymetrix Santa Clara, CA, USA GeneChip technology, microarray contract services, complete
microarray systems
Alphagene Woburn, MA, USA AlphaGenomics, full-length cDNAs, microarray contract services
Amersham Amersham, UK CyDye fluorescent-labelling reagents
Biodot Irvine, CA, USA Ink-jetting technology, microarray instrumentation
CLONTECH Labs Palo Alto, CA, USA Technology Access Program, gene-expression reagents
General Scanning Watertown, MA, USA Confocal-scanning instrumentation
Genetix Dorset, UK Microspotting instrumentation
Genome Systems St Louis, MO, USA Expressed-sequence-tag (EST) libraries
Genometrix The Woodlands, TX, USA Microarray technology platform, contract services
Genomic Instrumentation Services Menlo Park, CA, USA Development Partners Program, microarray instrumentation
Hewlett-Packard Palo Alto, CA, USA GeneArray, confocal-scanning instruments (Affymetrix)
Hyseq Sunnyvale, CA, USA HyChip products, genomics platform, contract services
Incyte Pharmaceuticals Palo Alto, CA, USA LifeSeq database, GeneJet and GEM technology, microarray
contract services
Intelligent Automation Systems Cambridge, MA, USA Custom automation, microarray instrumentation, contract work
Life Technologies Gaithersburg, MD, USA Fluorescent-labelling reagents
Molecular Applications Group Palo Alto, CA, USA GeneMine Pro, data-analysis and -visualization software
Molecular Dynamics Sunnyvale, CA, USA Microarray Technology Access Program, complete microarray
systems
Nanogen San Diego, CA, USA APEX, electronic microarray technology, contract services
Norgren Systems Palo Alto, CA, USA CCD-based imaging, microspotting instrumentation
OncorMed Gaithersburg, MD, USA Cancer prognostics and diagnostics
Pangea Systems Oakland, CA, USA GeneWorld, data-mining, -analysis and -management software
Protogene Laboratories Palo Alto, CA, USA Ink-jetting technology, microarray contract services
Qiagen Hilden, Germany DNA- and RNA-purification systems
Research Genetics Huntsville, AL, USA GenePairs, primers and purified PCR products
Silicon Graphics Mountain View, CA, USA Computational hardware and software, data-visualization and
-mining tools
Synteni Fremont, CA, USA GEM technology, microarray contract services
TeleChem International San Jose, CA, USA ArrayIt, PCR purification systems, microspotting technology,
scanners

sequences currently in public databases. Although density of 10 000 spots cm2. Because ink jetting does
microspotting is unlikely ever to produce the densities not require direct surface contact, piezoelectric deliv-
of photolithography, improvements in mechanical- ery is theoretically amenable to very high throughput.
spotting technologies will eventually allow the auto- Improvements in sample loading and sample changing
mated production of chips containing 100 000 features should, coupled with the inherent high-density capa-
in an area of ~6.5 cm2. Because of the ease of use and bilities of this approach, eventually enable the manu-
affordability, microspotting may become the microarray facture of complex microarrays. Piezoelectric-based
technology of choice for the basic research laboratory. delivery of phosphoramidite reagents has recently been
used for the manufacture of high-density oligonu-
Ink jets cleotide microarrays21. The successful application of ink
A third group of microarray technologies, the drop- jetting in a gene-expression setting (Fig. 3) demon-
on-demand delivery approaches, provide another way strates the immediate utility of this technology for
to manufacture microarrays (Fig. 2c). The most genome analysis.
advanced of these approaches are adaptations of the ink-
jetting technologies1821, which utilize piezoeletric and Combinations
other forms of propulsion to transfer biochemical sub- In view of the growing interest in microarray tech-
stances from miniature nozzles to solid surfaces nology and its potential impact on drug development
(Fig. 2c). Similar to the microspotting approaches, and disease profiling, it is unrealistic to suggest that any
drop-on-demand technologies allow high-density single enabling technology will dominate this large and
gridding of virtually any biomolecule of interest, diverse industry. A more balanced view suggests that
including cDNAs, genomic DNAs, antibodies and each of the technologies described above, and perhaps
small molecules. Ink-jetting technology is being devel- others22,23, will be utilized for the purpose they per-
oped at several centres including Incyte Pharmaceuti- form best at and will assist collectively in the prolifer-
cals (Palo Alto, CA, USA) and Protogene (Palo Alto, ation of microarray assays. The burgeoning microarray
CA, USA). industry, complete with scientific, business, financial
Although ink jetting is not currently as robust as and legal components, will soon provide a complete
photolithography or microspotting, this approach has repertoire of technologies and services for the scientific
been used to prepare microarrays of single cDNAs at a community (Table 1).

304 TIBTECH JULY 1998 (VOL 16)

 REVIEWS

Figure 3
Gene-expression monitoring with an ink-jetted microarray. This is a fluorescent scan of a high-density microarray printed using a GeneJet
(Incyte). Piezojet delivery of 200-pl droplets provides a density of 2500 cDNA groups cm 2. Coupling of the cDNAs to the chip surface
occurs via a succinimidyl-ester-displacement reaction. Array elements, printed as adjacent 9 12 subgrids, correspond to human cDNAs
selected from the LifeSeq database (http://www.incyte.com/), which contains approximately three million annotated expressed sequence
tags. The fluorescent sample was prepared from cultured human THP-1 cells by biotin incorporation into antisense RNA, followed by
secondary labelling with Cy-5 conjugated streptavidin (Molecular Probes, Eugene, OR, USA). Fluorescent intensities, represented in a
pseudocolour scale, reflect gene-expression levels.

Applications By correlating changes in gene expression with spe-

In a similar way to the development of recombinant cific changes in physiology, it is possible to gain mecha-
DNA and the polymerase chain reaction (PCR), nistic insight into a broad range of biological processes
microarrays have a large number of applications that (Fig. 4). Variations in gene expression in the normal
will expand and diversify over time2426. At present, population may, coupled with clinical data, have prog-
expression monitoring appears to be one of the most nostic value by allowing correlations to be made
biologically informative applications of this new tech- between the presence of specific expression markers
nology16,1113. For a number of reasons, chips are well and disease susceptibility (Fig. 4). Altered expression
suited to gene-expression analysis1,27; one theoretical patterns are expected to accompany many or all human
advantage of the use of micrarrays for expression assays diseases; thus, microarray analysis of diseased tissues
over some other applications is that these assays focus is expected to provide a wealth of diagnostic and
on the functional (expressed) segments of the genome. pathology information and yield potential drug targets
This aspect is important in complex systems, such as as well (Fig. 4). A gene whose expression level is dra-
the human genome, in which the ratio of coding to matically altered in a specific disease may provide a spe-
noncoding DNA is low28. Because expressed sequences cific target for small-molecule inhibition. The treat-
account for only ~3% of the genome, hybridization- ment of cells with small molecules has been shown
based transcript analysis effectively reduces the
complexity of the human genome by ~30-fold.
Microarrays of cDNAs provide expression infor- Inputs Outputs
mation for each gene represented on the chip. A
microarray of 100 000 unique human cDNAs should Normal tissue Prognosis
therefore allow the expression monitoring of the entire Diagnosis
human genome in a single hybridization. The enormous
information content of expression assays can be seen Diseased tissue Pathology
Microarrays
clearly when compared with chips designed for other (gene expression) Drug targets
applications, such DNA resequencing: resequencing
microarrays containing 100 000 groups of oligo- Drug efficacy
Test compounds
nucleotides would allow sequence checking for 25 000 Toxicology
bases, or approximately 0.0008% of the human genome.
Another advantage of the expression applications is
that the data serve as a direct link to function. Steady- Figure 4
state transcript levels provide a sensitive, global readout Microarrays for gene-expression analysis provide an integrated platform for functional
of the physiological state of a cells and tissue samples. genomics. Changes in the physiological state of the cells and tissues used for microar-
This is evident in the fact that specific patterns of gene ray analysis lead to specific changes in gene-expression patterns. Messenger RNA
expression have been observed as function of tissue from samples of interest (inputs) is isolated, labelled and analysed by hybridization-
type2, heat shock and phorbol-ester treatment3, and a based microarray analysis, yielding quantitative expression information for thousands
spectrum of metabolic and disease states5,6,1113; more- of cellular genes. Expression data, coupled with other types of information (such
over, these patterns are biologically informative and as that obtained in clinical and pharmacological studies) can have an impact on a
provide direct clues to gene function. spectrum of research areas (outputs).

TIBTECH JULY 1998 (VOL 16) 305

REVIEWS

to elicit specific changes in gene expression by itself3,4,6; Science 270, 467470

thus, because of the costly nature of clinical trials, 3 Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P. O. and
Davis, R. W. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 1061410619
microarray assays could provide a relatively inexpensive
4 Lockhart, D. J. et al. (1996) Nat. Biotechnol. 14, 16751680
platform for drug screening and toxicology studies 5 DeRisi, J. et al. (1996) Nat. Genet. 14, 457460
(Fig. 4). 6 Heller, R. A. et al. (1997) Proc. Natl. Acad. Sci. U. S. A. 94,
21502155
Conclusions 7 Fodor, S. P. A., Read, J. L., Pirrung, M. C., Stryer, L.,
The genome age will change biology forever, pro- Tsai Lu, A. and Solas, D. (1991) Science 251, 767773
viding sequence blueprints for numerous bacteria, 8 Khrapko, K. R. et al. (1991) Mol. Biol. 25, 581591
fungi, plants and animals. Whole genome sequences, 9 Maskos, U. and Southern, E. M. (1992) Nucleic Acids Res. 20,
the holy grail of stuctural genomics, will pave the way 16791684
for functional genomics by providing the information 10 Lennon, G. G. and Lehrach, H. (1991) Trends Genet. 7,
314317
required for microarray manufacture. Genome chips
11 DeRisi, J. L., Iyer, V. R. and Brown, P. O. (1997) Science 278,
will be a focal point of functional genomics by allow- 680686
ing the massive parallel analysis of expression patterns, 12 Lashkari, D. et al. (1997) Proc. Natl. Acad. Sci. U. S. A. 94,
mutations and many other types of genomic infor- 1305713062
mation2933. The data flood generated by chip-based 13 Wodicka, L., Dong, H., Mittmann, M., Ho, M-H. and
assays will be managed by powerful bioinformatics tools Lockhart, D. J. (1997) Nat. Biotechnol. 15, 13591367
capable of relational data analysis. Computer queries of 14 Fodor, S. P. A. (1997) Science 277, 393395
gene-chip databases should eventually enable virtual 15 Kallioniemi, A. et al. (1992) Science 258, 818821
biomedical research. 16 Shalon, D., Smith, S. J. and Brown, P. O. (1996) Genome Res. 6,
639645
17 Chee, M. et al. (1996) Science 274, 610614
Acknowledgments
18 Brennan, T. M. (1995) US Patent 5 474 796
We are grateful to members of the Stanford Bio- 19 Baldeschwieler, J. D., Gamble, R. C. and Theriault, T. P. (1995)
chemisty Department for their critical reading of the World patent application PCT WO 95/25116
manuscript and to the microarray technology group at 20 Blanchard, A. P., Kaiser, R. J. and Hood, L. E. (1996) Biosens.
Incyte including J. Blanchard, S. Park, T. Johann, Bioelectron. 11, 687690
S. Winder, J. Cormack, J. McDunn, T. Flynn, 21 Blanchard, A. (1998) in Genetic Engineering, Principles and Methods
R. Hardy, R. Gamble, J. Baldeschwieler, J. Lane, (Vol. 20) (Setlow, J., ed.), pp. 111124, Plenum Press
J. Snable, B. Koukis, S. Stuart, B. Cocks, G. Porter, 22 Eggers, M. et al. (1994) BioTechniques 17, 516525
L. Stuve, B. Wang, E. Goka, A. Chung, J. Au-Young, 23 Sosnowski, R. G., Tu, E., Butler, W. F., OConnell, J. P. and
D. Peterson, P. Hawkins, O. Bandman, R. Coleman, Heller, M. J. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 11191123
24 Shoemaker, D. D., Lashkari, D. A., Morris, D., Mittmann, M. and
L. Lin, T. Goralski, J. Ju, L. Ito, C. Silk, M. Doctolero,
Davis, R. W. (1996) Nat. Genet. 14, 450456
R. Nuttall, T. Carabeo, S. Lincoln, T. Maslyn, 25 Cronin, M. T. et al. (1996) Hum. Mutat. 7, 244255
M. Walker, B. Kennedy, T. Klingler, H. Hibbert, 26 Sapolsky, R. J. and Lipshutz, R. J. (1996) Genomics 33, 445456
R. Kopnisky, G. Bartha, J. Scott and J. Klemm for their 27 Schena, M. (1996) BioEssays 18, 427431
technical assistance. 28 Rowen, L., Mahairas, G. and Hood, L. (1997) Science 278, 605607
29 Lander, E. S. (1996) Science 274, 536539
References 30 Southern, E. M. (1996) Trends Genet. 12, 110115
1 Schena, M. and Davis, R.W. in PCR Methods Manual, Academic 31 Strauss, E. J. and Falkow, S. (1997) Science 276, 707712
Press (in press) 32 Hieter, P. and Boguski, M. (1997) Science 278, 601602
2 Schena, M., Shalon, D., Davis, R. W. and Brown, P. O. (1995) 33 Hoheisel, J. D. (1997) Trends Biotechnol. 15, 465469

Coming soon to Trends in Biotechnology:

Bioluminescent-bioreported integrated circuits form novel whole-cell biosensors

by Michael Simpson, Gary Sayler, Bruce Applegate, Steven Ripp and David Nivens

The impact of tw-hybrid and related methods on biotechnology

by Pierre Colas and Roger Brent

The development of site-specific delivery systems for biopharmaceuticals

by Dean Pettit and Wayne Gombotz

306 TIBTECH JULY 1998 (VOL 16)