Designing a Microarray Core Facility

Designing a Microarray Core Facility

By Mark W. Geraci, M.D.

University of Colorado Health Sciences Center

Denver, CO

The purpose of this segment is to detail the steps involved in designing and implementing a microarray core facility. Based largely on our experience at the University of Colorado Health Sciences Center, this segment details various aspects of a core facility that should be considered, and contains Internet links to the appropriate resources. There are six main areas to be considered when designing and implementing a core facility, and these include:

1. The choice of microarray platforms, including spotted arrays or Affymetrix as a platform.
2. The range of services to be provided by the core facility, including RNA preparation, labeling, hybridization and data analysis.
3. Providing data analysis software support.
4. Charges for services.
5. Providing continuing education for customers.

Microarray Platforms:

I. Spotted Microarrays

There has been an evolution in the technology and the substrate for spotted microarrays. Initially, most spotted arrays consisted of cDNA representations of between 100 and 700 base pairs and were pioneered by Pat Brown's laboratory at Stanford University. Subsequently, many companies and branches of the NIH sought to develop sequence-verified cDNA clone libraries. Many are commercially available for purchase from various agencies:

Incyte: http://www.incyte.com/expression/clones/index.shtml
Research Genetics: http://www.resgen.com/intro/clones.php3
National Institute of Aging: http://lgsun.grc.nia.nih.gov/

Difficulties with processing the cDNA clones and better technology for linking oligonucleotides to glass slides resulted in the emergence of custom oligonucleotide sets to represent cDNAs. The optimum length for oligonucleotide spotting remains to be determined, but certainly longer oligonucleotides (70mers) have more sensitivity and specificity than 40mers. Several oligonucleotide sets are available for spotting:

Incyte: http://www.incyte.com/expression/clones/index.shtml
Operon: http://www.operon.com/store/index
PGA Harvard: http://pga.mgh.harvard.edu/Parabiosys/resources/microarrays/
oligonucleotides/oligos_for_spotted_microarrays.php
MWG: http://www.mwg-biotech.com/html/discovery/
dna_arrays/oligo4sets.shtml

A. Advantages of Spotted Arrays

1. Flexibility of the microarray and the ability to customize or "direct" the microarray composition.
2. The cost of production of the microarrays can be less if produced locally.
3. The use of clone sets for follow-up studies.

B. Disadvantages of Spotted Arrays

1. Tremendous cost for equipment, including spotters, robots, and scanners.
2. Technical difficulties with spotted arrays and quality control for generating spotted arrays.
3. More difficult data sharing for spotted arrays, and the need to utilize RNA reference standards.

C. Equipment Needed

1. Spotters
   

   a. Quilled pin spotters
   http://www.genemachines.com/
   http://www.cartesiantech.com/
   http://arrayit.com/Products/Printing/printing.html
   http://www.biorobotics.com/ http://www.genetix.co.uk/
   http://www.genomicsolutions.com/
   http://www.robodesign.com/roboarrayer.htm
   b. Inkjet spotters
   http://www.cartesiantech.com/
   http://lifesciences.perkinelmer.com/areas
   /microarray/biochip.asp

   
   
   
2. Robotic
   Qiagen: http://www.qiageninstruments.com/instruments/
   Biomek: http://www.beckman.com/products/instrument/
   automatedsolutions/biomek/biomek2000_inst_dcr.asp
   
   
3. Scanners
   
   http://www.axon.com/GN_Genomics.html
   http://www.chem.agilent.com/scripts/pds.asp?lPage=398
   http://www.genomicsolutions.com/genomics/imaging.html
   http://www.virtekbiotech.com/product-chipreader.asp
   http://lifesciences.perkinelmer.com/areas/microarray/scanex.asp

II. Affymetrix:
http://www.affymetrix.com/index.affx

Currently, Affymetrix produces oligonucleotide microarrays encompassing a wide range of organisms and providing a variety of functions. These arrays include expression analysis for the following organisms: human, murine, rat, E. coli, Arabidopsis, yeast, Drosophila. Furthermore, genotyping can be performed using SNP analysis. This oligonucleotide platform has a variety of potential functions, and has more recently become commonly used.

A. Advantages of Affymetrix Arrays

1. There are multiple arrays available, as previously described.
2. Data between experiments can be normalized and used for comparison. Large expression databases can be generated. Some of these databases are commercially available, such as with GeneLogic: http://www.genelogic.com/
3. Equipment, scanners, and data analysis tools are readily available.

B. Disadvantages of Affymetrix Arrays

1. The arrays are somewhat expensive, and pricing is based on volume of use.
2. Changing of the microarray platform can sometimes be problematic and provide difficulty in databases that are built on previous platforms of Affymetrix arrays. Sources of assistance for platform comparison are becoming available from private sources as well as from Affymetrix (NetAffx)
   http://www.affymetrix.com/analysis/index.affx

C. Equipment

Affymetrix offers a wide variety of equipment, which largely centers on the GeneChip Instrument system. This system is all-inclusive and has the following components: the GeneArray® Scanner, the GeneChip Fluidics Station 400, the GeneChip Hybridization Oven 640, and a computer workstation loaded with the Affymetrix® Microarray Suite.

III. Services of a Microarray Core

One critical decision is the level of the services to be provided. Two different levels can be offered:

1. full RNA preparation and hybridization or,
2. customer instruction in proper RNA preparation and hybridization techniques; services include mainly scanning and data analysis.

A. Advantages

Full RNA services include the ability to minimize variability in experiments due to RNA preparation work.

B. Disadvantages

Having full-time workers involved in RNA preparation work is tremendously time-consuming. Most core facilities have moved towards a "wet lab" environment. In effect, customers bring samples (in various stages of preparation) to the laboratory. The laboratory is equipped with all the necessary reagents for complete RNA and hybridization work. These include such things as spectrophotometers, heating blocks, water baths, enzyme kits and labeling reagents.

IV. Data Analysis

There are a significant number of data analysis software products available on the market, and each core should have the availability of data analysis. The most commonly used (and user-friendly) software programs are enumerated with their respective websites.

1. Affymetrix provides a full set of data analysis software up to and including the Laboratory Information Management System (LIMS): http://www.affymetrix.com/products/software/index.affx
2. Silicon Genetics provides a widely used data analysis software algorithm called GeneSpring. This data analysis tool includes clustering algorithms, self-organizing maps, and discrimination protocols: http://www.silicongenetics.com/cgi/SiG.cgi/index.smf
3. BioDiscovery provides software encompassing image analysis for spotted arrays, array design for spotting arrays and data analysis for both Affymetrix and spotted arrays: http://www.biodiscovery.com/
4. Spotfire provides data analysis support for both platforms, and can be used for interactive "scripting": http://www.spotfire.com/
5. Molecular Mining provides several levels of data analysis support including GeneLinker Gold, which has traditional clustering and image analysis capabilities, and GeneLinker Platinum, with several higher-order discrimination capabilities: http://www.molecularmining.com/
6. We find PartekPro to be very useful for quick clustering and image analysis: http://www.partek.com/
7. Several publicly available analysis sites are available and these include:
   Eisen Lab: http://rana.lbl.gov/EisenSoftware.htm
   TIGR: http://www.tigr.org/software/
   Dchip: http://www.dchip.org/
   We have previously reported several other links in the following article.

V. Charges

Ideally, the core should be supported by institutional funds or core grants. These two sources will greatly reduce the cost for users, and therefore increase the availability of the expression analysis for scientists. Typically, there is a wide range of charges. Charges for complete "all-service" RNA preparation work range between $200 and $1000 per sample. The charge for hybridization and image analysis generally range between $75 and $300. These examples of charges are independent of the actual cost for the array, and only reflect the general charges for reagents and preparation of RNA.

VI. Continuing Education

Probably the most important aspect of a successful core operation is the ability to provide continuing education support for the customers. There are two aspects of continuing education that should be provided:

A. Education in Sample Preparation and Techniques

This can either be accomplished by on-line algorithms for RNA preparation or by printed materials to be made available to the customers.

B. Continuing Education for Data Analysis

Frequently there are changes in data analysis algorithms of commercially available microarrays or changes in the capabilities of commercially available software. Most companies offer education classes either on a national scale or, if large enough, on a local scale. We provide continuing education classes quarterly on these two topics and often invite representatives from industry to discuss their newest upgrades of product.

Several sites are available to provide updates on the technology of microarrays:

http://www.bioarraynews.com/index.htm
http://www.genomeweb.com/
http://www.bio-itworld.com/subscribe/
http://www.genpromag.com/scripts/default.asp
http://www.microarrays.org