- Open Access
Mining mouse microarray data
© BioMed Central Ltd 2001
- Published: 3 July 2001
Microarrays of mouse genes are now available from several sources, and they have so far given new insights into gene expression in embryonic development, regions of the brain and during apoptosis. Microarray data posted on the internet can be reanalyzed to study a range of questions.
- Embryonic Stem Cell Line
- Embryonic Lung
- Developmental Time Point
- Trophoblast Stem
- Trophoblast Stem Cell
The production of mouse microarrays has lagged somewhat behind the production of arrays from humans and from those model organisms for which full genome sequence is available, such as yeast or Caenorhabditis elegans. Oligonucleotide-based arrays from sources such as Affymetrix (Santa Clara, CA, USA) have been in production for a number of years, but their high cost has been a barrier to widespread general use in academia. The cDNA array effort has been hampered largely by the scarcity of large, well-annotated cDNA clone sets from mouse tissues. Fortunately, a number of recent publications have addressed this issue. Minoru Ko at the National Institute on Aging (NIA; National Institutes of Health, Bethesda, MD, USA) has developed a 15,247 clone set derived from mice largely at early developmental time points, with libraries covering stages from the early blastocyst to embryonic day (E) 7.5 (for which there are embryo and ectoplacental cone samples) . This set will have added to it later this year a further 11,000 clones derived from sequencing of similar libraries, and libraries from trophoblast stem cells, hematopoietic stem cells, embryonic stem cell line R1, newborn heart, brain, and kidney, and mesenchymal stem cells (M.S. Ko, personal communication). The set has been sequence-verified  and is cloned into a vector system with binding sites for both SP6 and T7 polymerases flanking the insert, so that riboprobes for in situ hybridization studies can be generated directly. The average insert size is estimated to be roughly 1.5 kilobases (and we have found this generally to be the case). The clone set has been widely distributed to a number of centers for microarray printing, including our own local facility at the Ontario Cancer Institute ; these centers are currently in various stages of production .
The RIKEN Institute in Japan has also recently described a 21,076 cDNA clone set thought to represent at least 12,890 unique genes, which was generated from sequencing just under one million total expressed sequence tags (ESTs) [6,7]. This group of clones has the major advantage that more than 60% are full-length cDNAs, because of the methods used in library construction. Also, these clones are derived from a broad spectrum of adult and developmental stages of various tissues, which provides a nice complement to the much earlier stages used to derive cDNA libraries for the NIA 15,247 clone set. A quick analysis indicates that there are at least 26,700 non-redundant clones between the two sets. Although arrays have been generated from these clones by RIKEN (see below), they have not yet been made widely available to the scientific community. As a result, we will probably see more work initially in the literature that uses microarrays printed from the NIA set in combination with custom-derived clones.
Arguably the most comprehensive mouse arrays available are the Murine Genome U74 Set from Affymetrix. This group of three arrays provides coverage of roughly 36,000 mouse genes, of which about 30,000 are known only as ESTs. As mentioned previously, however, there has not been widespread use of these arrays because of the prohibitively high cost of the arrays and the associated scanning hardware. But given the advantages of oligonucleotide-based arrays for interrogating multiple regions of a transcript or alternatively spliced forms, in addition to their flexibility for array design, we expect the adoption of this platform for array technology to increase in the future. Operon Technologies (Alameda, CA, USA) has recently introduced a mouse 7,000 oligonucleotide set that is available for purchase for array printing;the size of this and similar sets is likely to increase substantially in the near future.
In general, given the advances in size and availability of sets of mouse cDNA clones and oligonucleotides, we foresee an increase in the number of centers making the initial investment to start microarray facilities and capitalize on this technology. Printing cDNAs has the advantage of replicating directly the microarrays printed elsewhere, although it does have the disadvantage of laborious amplification of cDNA inserts for printing and difficulties in managing a large number of clones. Oligonucleotide arrays eliminate some of these problems and offer a number of flexibility advantages in array printing, but they come with higher initial start-up costs. Although we are moving in the direction of a 'whole-genome microarray' as we approach the completion of the mouse genome-sequencing project, it is likely that many groups will focus on smaller, custom, gene groupings to reduce both costs and the engineering difficulties associated with printing arrays of higher density.
Although to date only a small number of mouse microarray studies have been published in the literature, we expect a tidal wave of papers over the next two to three years as this technology becomes widely available. Minoru Ko's group has described experiments using nylon arrays printed from their NIA 15,247 clone set. In profiling comparisons of E12.5 placenta and whole embryos, they identified 720 genes that appeared (statistically, from triplicate measurements) to be upregulated in the placenta . Of these, 181 were represented by novel ESTs not previously known to be placentally enriched, increasing the total number of genes known to be highly expressed in the placenta by about five-fold. The placenta-specific expression pattern was confirmed for one of these novel genes (H3018Bo6) by in situ hybridization; homology searches suggest that it could be a novel member of the placental growth-related hormone family.
The RIKEN group has created cDNA microarrays printed on glass slides from a set of 18,000 clones enriched for full-length cDNAs. In a recent paper , gene expression was profiled in a total of 49 different adult and embryonic tissues. Data were normalized using RNA from E17.5 whole embryos, which has a number of practical advantages as an external reference over using individual tissues, given the ease of RNA preparation from whole embryos and the diversity of expressed genes. Data were then clustered  to look at the underlying order in gene expression amongst various tissues. As expected, the observed expression profiles correlated well with the tissues from which cDNAs were originally cloned. Further analysis of expression patterns in the developing nervous system showed unique clusters associated with apoptotic cell death in neural remodeling and with different phases of the cell cycle. In an analysis of expression patterns of metabolic pathways, a separation of enzymes into 78 different synthetic and degradative metabolic pathways was used to look at coordinated regulation of expression patterns . Interestingly, genes were clustered into groups of those ubiquitously expressed and those with tissue-specific patterns. In a more detailed analysis of enzymes in the glycolytic pathway, tissue-specific differences were identified in gene products for glycolytic enzymes derived from muscle versus those present in liver and kidney .
Using the Affymetrix system, Zirlinger et al.  have recently analyzed differential gene expression in selected regions of the brain using arrays containing approximately 34,000 genes. They screened by in situ hybridization a series of 33 genes identified to be amygdala-enriched from array experiments and found that a number of them had expression patterns restricted to the subnuclei of the amygdala. Using the earlier Mu11K array, containing 11,000 gene elements, Voehringer et al.  demonstrated a series of alterations in gene expression in apoptosis-sensitive compared with apoptosis-resistant B-cell lymphoma cell lines before and after irradiation. From these observations, they suggested a model for sensitivity to apoptosis involving induction of transcripts for genes participating in mitochondrial uncoupling and loss of membrane potential.
One of the major stumbling blocks for new users is the processing and management of data once microarray hybridization has been done. The number of both freely available and commercial packages that are user-friendly for performing these functions is rapidly increasing. At our institute, we use a combination of the commercial Quantarray (Packard BioChip Technologies, Billerica, MA, USA) and GenePix (Axon Instruments, Union City, CA, USA) packages to find spots and assign grayscale hybridization intensities. We then use a package developed locally  to handle background subtractions and normalizations and to create databases of cumulative experiments for clustering. These are certainly not the only tools available, nor are they necessarily the best, and we anticipate more user-friendly packages for these functions to arrive shortly, given the flurry of activity in the area. What is not yet clear is how best to sort through the data from validated experiments once one has identified groups of genes that are of interest to the individual investigator.
We expect the volume of microarray work performed in the mouse to quickly outweigh that of other model organisms for a number of reasons. From a developmental perspective, the mouse is the only genetically tractable model organism that closely resembles humans in the formation of organ systems with similar structure and function to ours. Collecting fresh tissue for RNA analysis at all developmental and adult stages is a relatively simple task. Also, the number of genetic mutants closely resembling human disease processes is dramatically increasing with the expansion of forward-genetic (mutagenesis) screens in the mouse, providing many opportunities to perform genome-wide scans of transcripts varying in a particular model of interest. Wide application of microarray techniques to embryos, tissues, or cells harboring engineered mutations will create an exponential number of permutations and combinations ripe for investigation.
Although the future is bright, there is room for improvement in a number of areas. A working draft copy of the mouse genome sequence will certainly aid in making sense of expression data from novel ESTs. This could potentially narrow down genes of interest to those that have a putative motif or predicted function of interest. As judged by discussion at a recent symposium on statistical aspects of microarray data analysis , there is not yet anything resembling consensus amongst statisticians as to the most appropriate methods to deal with array data. One of the overall challenges to statisticians and computer scientists in general will be the development of analysis software that is readily usable by the uninitiated biologist with data from their first experiment in hand. Although many companies currently claim that their products do this and everything else, this has not been the case in our experience. This will no doubt change in the near future, however, given the rapid advancements from academic and commercial activity in the area.
The authors thank Hidemasa Bono and the RIKEN Hayashizaki mouse genomics group for sharing data from their microarray database. We also thank Chi-Yip Ho and Tilo Kunath for many helpful discussions.
- Schena M, Shalon D, Davis RW, Brown PO: Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. 1995, 270: 467-470.PubMedView ArticleGoogle Scholar
- Ko MS, Kitchen JR, Wang X, Threat TA, Wang X, Hasegawa A, Sun T, Grahovac MJ, Kargul GJ, Lim MK, et al: Large-scale cDNA analysis reveals phased gene expression patterns during preimplantation mouse development. Development. 2000, 127: 1737-1749.PubMedGoogle Scholar
- Kargul GJ, Dudekula DB, Qian Y, Lim MK, Jaradat SA, Tanaka TS, Carter MG, Ko MS: Verification and initial annotation of the NIA mouse 15K cDNA clone set. Nat Genet. 2001, 28: 17-18. 10.1038/88206.PubMedGoogle Scholar
- Microarray Center at University Health Network, Ontario Cancer Institute. [http://www.oci.utoronto.ca/services/microarray]
- NIA/NIH Mouse Genomics Home Page. [http://lgsun.grc.nia.nih.gov]
- The RIKEN Genome Exploration Research Group Phase II Team and the FANTOM Consortium: Functional annotation of a full-length mouse cDNA collection. Nature. 2001, 409: 685-690. 10.1038/35055500.View ArticleGoogle Scholar
- RIKEN Genomic Sciences Center, Genome Exploration Research Group. [http://genome.gsc.riken.go.jp/index.html]
- Tanaka TS, Jaradat SA, Lim MK, Kargul GJ, Wang X, Grahovac MJ, Pantano S, Sano Y, Piao Y, Nagaraja R, et al: Genome-wide expression profiling of mid-gestation placenta and embryo using a 15,000 mouse developmental cDNA microarray. Proc Nat Acad Sci USA. 2000, 97: 9127-9132. 10.1073/pnas.97.16.9127.PubMedView ArticleGoogle Scholar
- Miki R, Kadota K, Bono H, Mizuno Y, Tomaru Y, Carninci P, Itoh M, Shibata K, Kawai J, Konno H, et al: Delineating developmental and metabolic pathways in vivo by expression profiling using the RIKEN set of 18,816 full-length enriched mouse cDNA arrays. Proc Natl Acad Sci USA. 2001, 98: 2199-2204. 10.1073/pnas.041605498.PubMedView ArticleGoogle Scholar
- Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998, 95: 14863-14868. 10.1073/pnas.95.25.14863.PubMedView ArticleGoogle Scholar
- Zirlinger M, Kreiman G, Anderson DJ: Amygdala-enriched genes identified by microarray technology are restricted to specific amygdaloid subnuclei. Proc Natl Acad Sci USA. 2001, 98: 5270-5275. 10.1073/pnas.091094698.PubMedView ArticleGoogle Scholar
- Voehringer DW, Hirschberg DL, Xiao J, Lu Q, Roederer M, Lock CB, Herzenberg LA, Steinman L, Herzenberg LA: Gene microarray identification of redox and mitochondrial elements that control resistance or sensitivity to apoptosis. Proc Natl Acad Sci USA. 2000, 97: 2680-2685. 10.1073/pnas.97.6.2680.PubMedView ArticleGoogle Scholar
- Tyers M, Jorgenson P: AFM 4.0: A toolbox for DNA microarray analysis. Genome Biology. 2001,Google Scholar
- Biodiscovery Symposium: Experimental Design and statistical Analysis in Array Technology. [http://biodiscovery.com/services/ws5.html]