AT A GLANCE
•Director of the General Clinical Research Center’s Core Molecular Biology and Genomics Laboratory at UCSF/San Francisco General Hospital;
•Co-Chair of the Association of Biomolecular Research Facilities (ABRF) microarray research group;
•Co-director of the recently established University of California San Francisco DNA Bank facility;
•Administrator of the Gene-Arrays Listserv.
Received a masters degree in cell & molecular biology from San Francisco State University in 1987.
Interests: nature macrophotography, film/cinematography. Passion — music.
QWhat role do microarrays play in research at UCSF?
AArrays have been primarily used for human and murine expression profiling, though we have used yeast arrays. The array studies to date have been in areas such as dermatology, stomatology, HIV, and prions.
QHow is the microarray facility set up within the university?
AThe clients are responsible for supplying the arrays and a final IVT sample. The core handles the last stages of sample preparation, hybridization, washing and staining, scanning, and preliminary evaluation of the quality of the data. All samples are first evaluated via test arrays, and if all proves well, they are then hybridized against the expression arrays. Currently, the primary data is burned onto CDs, and the client receives one copy and the core retains the other in case the first disk goes missing.
QWhat types of microarrays do you use?
ACurrently we are only using the Affymetrix microarray platform. Our primary expression arrays are the human U95A and murine U74A arrays. We’ve also used p53 and CYP450 arrays, and we have a number of studies in the works involving newer arrays such as one for expression profiling of Candida.
QWhat methods do you use to analyze microarray data?
AFor lack of a bioinformatics staff, analysis is limited to an evaluation of the quality of the data. However, we’ve recently acquired a new Data workstation, which I hope will allow me to provide a generic ‘first pass’ look at the data. It would be nice to be able to supply every client with a preliminary data set that presents a conclusion or an overview in biological terms and not a seemingly endless spreadsheet that looks more like an accounting nightmare than a biological experiment.
QWhat is the biggest challenge you face in working with microarrays?
AIt would be nice to see a bit more forthrightness on a number of issues from the system’s vendor [Affymetrix]. A good amount of time and effort has been wasted sorting out issues related to the platform. Also, one issue I have found to be somewhat delicate is that of clarifying just how involved the data analysis is, and how readily one can ‘arrive’ at conclusions that in fact are questionable at best. Without care one can be readily misled by a misinterpretation of the data.
QHow do you tackle these challenges?
AOn the issue of the data analysis, one of the first things I do try to do is immediately gauge a researcher’s expectations and knowledge of microarrays. If there is some question in their minds as to what they can expect from the study, I try then to make as clear as possible just what questions the technology can address and to what extent. Often I think that the whole microarray technology has been oversold in many respects as to just what it can do, or what insights it can provide.
QIf you could make out a wish list for microarray technology advances or improvements over the next couple of years, what do you most want or need?
AI would like to see the hybridization times for the arrays sharply decreased. One possibility may come from advances in the microelectronics arena. That would dispense with the overnight hybridization steps now common. Scanning is also a limitation at this point. In the case of Affymetrix, one can process a good number of arrays using multiple fluidics stations, but given that most labs have but one scanner, a bottleneck is unavoidable.
Like most, I would also very much like to see a dramatic decrease in the sample requirements, particularly in the case of human studies where samples are precious and generally very small. A number of studies are not currently possible if for no other reason than the limitations imposed by the current sample requirements.