Microarray Core Director
Mark Barton Frank, Ph.D.
Assistant Research Member, Arthritis and Immunology Research Program
Oklahoma Medical Research Foundation
825 NE 13th Street
Oklahoma City, OK 73103
User Fees Core Services Core Profile Analysis Methods Microarray Production
The Oklahoma Microarray network links hitherto independent functional genomics groups at the University of Oklahoma, the Oklahoma Medical Research Foundation, the University of Health Sciences Center and Oklahoma State University and the University of Tulsa. The Oklahoma Microarray Network provides all participating molecular and cellular laboratories with functional genomic technology, allowing them to incorporate genome-scaled thinking into their research. BRIN funding has formalized the Oklahoma Microarray Network to make this technology more broadly available to Network researchers, foster further technological innovation in the general area of functional genomics and bioinformatics, and ultimately improve the competitiveness of Network participants for NIH funding.
The services provided by the Microarray facility include consultation on experimental design, RNA QC and labeling, microarray printing, hybridization, image analysis, and bioinformatics. The scientific team plays an active role in the evolution of the sciences of microarray production and processing, bioinformatics, and microarray database design. INBRE funds are being used to support a portion of the facility’s personnel, which is currently staffed by 19 multidisciplinary scientists.
Microarray Core Services
 | Consultation on experimental design and RNA preparation |
|
RNA analysis |
RNA quality via capillary gel electrophoresis -Agilent 2100 Bioanalyzer
RNA quantity via spectrophotometer
Help users optimize RNA isolation procedures to optimize yield and quality
|
Labeling |
Preparation of labeled cDNA
High throughput 96-well plate-based labeling used. Label up to 96-samples in parallel (high throughput, reduced variation)
|
Hybridization |
Ventana Discovery hybridization station (20 samples/run)
Automated rehydration, denaturation, hybridization, & washing
|
Scanning and Quantitation |
Agilent dual laser scanner, 5 micron resolution, 48 slide carousel
Imagene (Biodiscovery) quantitation software
|
Bioinformatics analysis. Includes necessary custom methods development. |
|
Validation |
Quantitative real-time PCR
Assistance preparing manuscripts and grants
The facility has expanded three times to its current size of 4,000 square feet to meet the growing needs of its users. The facility has steadily increased gene density (printed slides are now genome-scale) and sample processing throughput (currently 20 microarrays per day). Within the first 8 1/2 months of 2003, over 1,100 microarrays were printed, processed, and analyzed for 31 research groups. The OMRF Microarray Core facility has developed the requisite expertise to be the primary microarray core for the state of Oklahoma, and is fully capable of supporting microarray efforts of the satellite campuses.
The bioinformatics group has 5 members, 3 bioinformaticians, a database engineer, and a biostastician. Working with the INBRE Bioinforamtics Core, these individuals provide for the analytical needs of the Microarray Core users. The methods used by the facility were developed by this group and include methods for normalization, differential gene expression analysis, clustering, and gene networking (identifying interacting gene sets). Unlike commercially available microarray analysis software and the majority of freeware, these in-house developed methods establish statistically relevant thresholds such that sensitivity and specificity are maximized. These methods are reviewed by the bioinformatics community and are becoming part of the standard in this evolving field. Most importantly, this bioinformatics group has taken the lead in developing the BASE (BioArray Software Environment) accepted standard microarray database, used by over one hundred microarray labs world-wide. This group is a principal developer of BASE in conjunction with the BASE project group at Lund University, responsible for BASE documentation, collecting software bugs and bug fixes, user support, and future updates. The next release, BASE 1.4, will reduce installation complexity and improve operating system compatibility. We have also developed BASE documentation and run the website for reporting bugs or for contributing code for program revision.
Computational Support. The OMRF facility currently uses a Dell Poweredge 2650 dual-processor server with a RAID 5 array for primary storage. The lab's fileserver is a 16-bay hot-swap drive array, capable of staged expansion. Current storage capacity is 800 gigabytes, with the possibility of expanding out to 2.3 terabytes using currently available hard drive technology. This system is also protected by a RAID 5 array. The fileserver backs up over a secure VPN connection to an offsite 20-bay hard drive array, updated 12 times a day, with a snapshot every 24 hours. Current usage patterns allow archiving of 2 years of snapshots, which could then be moved to tape, The OMRF Microarray Core Facility also utilizes a PC network with fifteen workstations running Microsoft Windows 2000 and 3 Powermac G4s running OS 10.2.
Microarray Production and Processing
In collaboration with Qiagen-Operon Inc., the OMRF Microarray Facility has developed processes for producing genome-scale oligonucleotide-based microarrays. When this work began approximately 2.5 years ago, no reliable method for oligonucleotide attachment existed. These methods are now the standard method issued by Qiagen-Operon with their libraries, and the OMRF Microarray core is a reference lab for their continued development. The microarrays utilize 70 base pair long DNA oligonucleotides with optimized length and sequence specificity, reducing or eliminating the cross-hybridization problems encountered with cDNA-based arrays. The human arrays have 21,329 human genes represented and the mouse arrays have 16,463 mouse genes represented. The oligonucleotides were derived from the UniGene and RefSeq databases (http://ww.ncbi.nlm.nih.gov). This library is among the most comprehensive collections of human genes available on any array format. A complete listing of the genes can be seen at http://www.omrf.ouhsc.edu/~frank/microarray.html.
Microarray Data Analysis Methods
All
of the OMRF Microarray Core custom bioinformatics programs were written using
MatLab and Statistica statistical software and are available on request from igor-dozmorov@omrf.ouhsc.edu.
These tols allow for normalization of microarray data, definition of an internal
reference group for differential expression analysis, identification of genes
differentially expressed in patients vs. control group by processes that we term
Associate Methods, discriminant function analysis (DFA), and biochemical
function and pathway analysis. The OMRF Microarray core facility is using
BASE for databasing, Matlab, and Statistica for developing computational tools,
and Spotfire for its user friendly graphical user interface. A series of
libraries and data exchange scripts (denoted wrapper scripts) supporting the
connection of these programs is being developed in the languages Perl and
Python. To all the end user access to the analytical tools through web
interface, a downstream user interface is being created. To ease the
incorporation of new anlysis methodologies being developed, the libraries and
wrappers are flexible enough to be used with minor modicification by any
chainable process. The experimental methods that will be employed to
streamline the bioinformatics are outlined below.
· Preliminary Simple Object Access Protocol (SOAP) wrapper scripts
· Basic text-stream protocols established
· Create libraries for producers for basic text-stream protocol
· Outline of advanced text-streams
· Move entire current analysis over to stream-based design
· Design of support infrastructure for publishing objects and naming protocols
The OMRF Microarray Facility is working towards self-sufficiency by obtaining users fees for services. The core provides the following steps for users: consultation on experimental design, RNA QC and labeling, microarray printing, hybridization, image analysis, and bioinformatics. This service costs $750/sample including personnel, reagent, and equipment upkeep. These costs are passed on to those with grant support for this technology. The current throughput of the facility 80 samples/week is sufficient to generate the required operations revenue. Presently the volume of users is not adequate to completely fund operations. Our facility is supported in part by institutional funds to compensate for this shortfall. Institutional funds provide a means of processing samples for investigators to generate preliminary data that is then used to obtain grant funding. Several grants have been obtained by users with this data and thereby increased our user base. In addition, we have established collaborations with investigators outside of Oklahoma to increase the throughput in the facility and help fund operations. Finally, our facility is generating publications, 5 published or in press to date. All of these publications are recent, (all published in 2003). We have seen that the interests in the facility grow since it has proven capable of generating publishable data and it is likely that our continued success will further expand our user base to help cover operations costs.