At A Glance
Completed postdoctoral research at the National Cancer Institute Frederick Cancer Center investigating the mouse oncogene.
Joined the National Institute of Environmental Health and Safety (NIEHS) in 1990. After NIEHS researcher Cindy Afshari opened the institute’s microarray facility in 1998, joined Afshari as co-director and toxicogenomics facilitator.
Research interests include investigating DNA damage responses focusing on the role of cell cycle checkpoint response genes, the ATM and ATR genes, in dealing with DNA damage.
Received his PhD in experimental pathology from University of North Carolina.
QHow did you and Dr. Afshari go about setting up the microarray facility at NIEHS?
AWe worked with Jeff Trent’s lab at the National Human Genome Research Institute (NHGRI), and tried to recreate their effort down here. We had someone build an arraying robot based on the Pat Brown arraying technology, and started printing cDNAs ourselves.
QDo you still use the Pat Brown model robot?
AYes, but we have modified it in various ways. It uses Telechem titanium quill pins. We spot onto poly(L)-lysine glass slides, and spot under humidity-controlled conditions. We have found that humidity is very important for regulating spotting. We have also built in a wash station so we get uniformity in spots. Currently we are moving from 96- to 384-well plates.
QWhat do you spot on your arrays?
AWe have cDNA clones from human, mouse, rat, and Xenopus laevis (the African clawed frog, whose embryos are used in toxicology studies), and also have complete coverage of the yeast genome. Dr. Perry Blackshear has been gathering Xenopus clones, and is driving that effort. The mouse clones are IMAGE clones from Incyte, the rat clones and some of the human clones are from Research Genetics, and some human clones come from our collaboration with NHGRI.
A universally recognized problem with cDNA is that often the sequences are not what they say they are. So if someone is interested in a particular genetic effect or biomarker, we have the clone resequenced, and we validate microarray gene expression profiles using Northerns and TaqMan. We also feel that it’s important to do replicates. We usually do four replicates, consisting of two fluor flips: For each pair of replicates, we reverse the Cy3 and Cy5 dyes. Replkicates help us to differentiate real expression vs. random chance.
QWould you consider going with a prefabricated oligo chip?
AWe like having flexibility in our chip design. The first chip we printed was a 2,000-gene human toxicology chip. We have also attempted to get total genome coverage. So far we have 7,000 rat clones on a chip, almost 9,000 mouse clones, and have printed up to 18,000 clones on a human chip. We are looking at spotting oligonucleotides, because they would give us some additional flexibility.
QHow do you process and analyze data?
AWe have two Axon scanners, and just brought on an Agilent scanner that has a carousel loading mechanism. We have a bioengineer who has written our own script for the Axon scanners that balances intensities so we get equal distribution of intensities in both Cy3 and Cy5 channels.
For image analysis, Yidong Chen has written a robust program, which he is marketing through Scanalytics. It finds the pixels and the spots, and can handle irregular spots. It also subtracts background intensities in each channel for each spot, and will normalize and give you the ratios of spot intensities. It allows you to evaluate the distribution of differentially expressed genes on a statistical basis. In this way, it allows us to identify the outliers based on a 99 or 95 percent confidence interval, not on a twofold or threefold cutoff.
We also have a very good bioinformatics team, which has developed software that incorporates cluster analysis, single gene ANOVA, linear discriminants, principal component analysis, and other analysis tools, some parametric and some non-parametric. It’s very important that you bring in a number of tools because at this point there’s no general agreement on issues such as normalization, and the best approach to analysis. We are very open to developments in the field and we are trying to keep an open mind.