AT A GLANCE
MD PhD, University of Turku, Finland. Worked as a surgeon and a general practitioner, then did PhD research on regulation of syndecan-1 proteoglycan.
Joined Peter Ratcliffe’s group at Oxford as a research fellow studying cellular behavior in conditions of reduced oxygen availability (hypoxia).
Current research focuses on how oxygen availability regulates gene expression, on the role of individual components of the oxygen sensing system in gene regulation, and their interplay with two other fundamental metabolism regulators, iron and glucose.
QCan you tell me a little bit about the role that microarrays play within the research aims of the Turku Center for Biotechnology?
ATurku Center for Biotechnology has over ten independent research groups, many of which study diverse aspects of cell signaling. The microarray center was established about two years ago with an aim to provide the whole country with diverse types of DNA chips. The center has built up spotting, EST library, and PCR handling robots, as well as supporting software packages. In a small country such as Finland it is much more efficient to concentrate the production of chips in one place rather than having spotting facilities at each university or institute.
QWhat kinds of chips, arraying equipment, scanners, and analysis software do you have?
ACurrently we have two spotters — a custom made spotter and a Virtek ChipWriter Pro — and four robots taking care of plasmid purification, PCR, slide coating, consolidation, etc. We produce both “ready made-chips” currently consisting of between 3,000 genes (with three replicates for each gene on a slide) and 10,000 genes (with one cDNA per gene) as well as custom chips where customers can select individual genes from our gene lists. We’ve used the human Research Genetics 40K library, the [National Institute on Aging] mouse 15K library, and an Arabidopsis library. We also host and run an Affymetrix platform.
QHow do you QC your chips and what kind of throughput do you have?
AWe have integrated 48 different spots at different DNA concentrations for data normalization and hybridization validation. Each slide is bar-coded and we have developed a LIMS database where we enter production and other parameters for every slide and spot, which can be traced back after hybridization. We include gel images together with DNA quantity and length of each PCR product, gene sequences, DNA stainings, and hybridization images for slide batches, as well as physical production parameters such as spotter room humidity. Currently we can produce from 1,500 to 4,000 chips per month.
QIs equipment service a particular issue in Finland?
AYes. We’d like to be able to service the machines ourselves as much as possible. Many times we simply cannot afford to wait a long time for mostly very expensive service to reach us from the other side of the world.
QWhat is your biggest challenge with microarray technology, and how do you address it?
AOne of the main problems we have faced is the quality of EST libraries. They contain poorly growing or empty clones, contaminated clones, and mistakenly annotated sequences. To overcome this problem (and avoid plasmid purification and PCR production), we are currently testing 70-mer oligonucleotide arrays.
QWhat methods do you use to analyze microarray data, and what is the biggest microarray analysis pitfall?
AWe use the [Packard] ScanArray 5000 Scanner and QuantArray software together with Kensington data mining software. Bioinformaticians and mathematicians at the center also make clustering and statistical tools as well as [inductive logic programming] and artificial neural networks available for data mining. I feel that the major analysis pitfall is the biological variation (due to circadian variation, small temperature changes, changes in pH, etc.) as well as other “unexplainable” variations between independent controls or superficially similar samples. Too many publications seem to ignore these.
QIn the world of microarrays, what developments would you most like to see emerge in the coming year or so?
AProper protein arrays. If we ever want to understand the networks operating within cells and between cells, we will need to understand the global pattern of protein interactions in diverse situations.