A Different Kind of Clustering
Do genes expressed within the same tissue — say placenta or skeletal muscle — clump together along the genome?
Although two relatively recent papers on genome-wide maps of these tissues suggest this is the case, a new study published online in Nature Genetics last week provides contrary evidence.
Martin Lercher and colleagues at the University of Bath department of biology and biochemistry examined serial analysis of gene expression data for 14 tissues, looking for clusters of genes along the genome (not to be confused with co-expressed groups produced by clustering algorithms). They removed tandem duplicates, and then, after finding clusters, confirmed these clusters using expressed sequence tag data.
What they found is that expressed genes do not map to clusters on a genome, except when they are housekeeping genes.
However, the group did not rule out “special cases of clustering within tissues. “It has been suggested that clustering may be selectively favored because it allows linkage disequilibrium to more easily be maintained,” they wrote. “Imprinting, such as that of placentally expressed genes, may provide another mechanism underlying tissue-specific clustering.”
GenMAPP Helps Researchers Find Their Pathway
Up, down, expressed all around — the gene expression data is only so informative if you can fit the expression patterns into the overall metabolic mechanism of the cell. This is why Bruce Conklin at the Gladstone Institute developed GenMAPP (Gene MicroArray Pathway Profiler), a freely available Windows-based tool that graphically represents gene expression values on pathway diagrams.
“Clustering tools take an unbiased view and assume you know nothing about biology, and for many genes that’s exactly what we know — nothing,” said Conklin. But if you know a lot about the genes, you want to take a totally biased view of your genomic experiment in the context of known biochemical pathways.”
The program starts with MAPP, a drawing component of the program that represents pathways with graphical objects such as genes, receptors, subcellular components, lines, and arrows that can be placed and manipulated on a “drafting board.” Users can then upload any set of gene expression data as a comma-separated file in order to color-code the gene objects to represent their degree of up- and down-regulation.
Conklin’s group has created a “starter set” of pathways based on textbooks, review articles, and public pathway databases. There are currently 1,009 MAPPs for mouse, 1,905 for human, 345 for rat, and 633 for yeast at www.genMAPP.org.
Users can modify the existing MAPPs or create their own, and can share MAPPs with colleagues. This capability gives Conklin hope that the system will eventually become the Adobe Acrobat of gene expression data — the standard format for sharing information in the field. He envisions researchers creating and sharing MAPPs and eventually including them in research papers so that readers can download the file and add their own gene expression data to the pathway.
Since GenMAPP was launched a year ago, over 1,300 researchers have registered to use the software, Conklin said. Around half of these are commercial users.
A paper on GenMAPP, which Conklin co-authored, appears in the May issue of Nature Genetics.
Illumina Gets $1M For BeadArray Development
Illumina of San Diego has been awarded a $1 million Phase 2 grant from the National Institutes of Health’s Small Business Innovation Research program to develop its BeadArray technology.
Illumina’s technology consists of fiber optic stalks with wells at the end of each fiber optic, and a microbead in each well. The beads are labeled with a fluorescent tag. There are 25 of each kind of labeled bead, making for 2,000 unique beads per bundle.
These bundles can be placed on 96-, 384-, or even 1,536-well plates, yielding a dizzying array of arrays. A 1,536-well plate holds more than 3 million uniquely labeled beads and a standard 96-well plate has 192,000 assays.
“This grant will allow us to develop new, automated systems for generating high-quality data from many samples using our parallel array format, stated Mark Chee, Illumina’s vice president of Genomics. “We believe this will enable researchers to accelerate their studies of genetic variation and function and help to catalyze the development of personalized medicine.”
The SBIR grant follows on the heels of two other collaborations that Illumina has initiated. Earlier in the month, Illumina signed an agreement with Placer, a technology company that is looking at SNPs associated with complex diseases, including psoriasis. At the end of April, Illumina signed an agreement to provide the Laboratory of Psychiatric Genomics at the University of California, San Diego, with SNP genotyping services.
Illumina and Applied Biosystems are expected to release their genotyping product by the end of June.
Casting a Wide Net For Affy Data
For Affymetrix users who would like to participate in the Affymetrix retrospective data study conducted by the Microarray Research Group of the Association of Biomolecular Resource Facilities, the study is more accessible than ever.
In addition to the ABRF website, http://www.abrf.org, the PDF files detailing the study can be downloaded from Berkeley: http://zenith.berkeley.edu/genearrays/Files/; France: www.DNAarray.com, or The Yahoo Microarray e-group: http://groups.yahoo.com/group/ microarray/files/Invite.pdf and http://groups.yahoo.com/ group/microarray/files/margFTP.pdf
The study, which is being headed up by Andrew Brooks of the University of Rochester Medical Center, is looking at sources of variability with Affymetrix arrays. Data submission began in early May.