Summary

In a typical experiment, a user would bring between 100 ng - 500 ng of total RNA for each array assay, and within about one week, receive analyzed expression data for 24,000-48,000 human or mouse genes. The IDM Shared Resource will assist with follow-up bioinformatic analysis, and interesting regulated gene targets can be rapidly validated by real-time PCR using the Q-PCR service.

Overview of human and mouse microarray analysis

The major service now offered is gene expression analysis using the Illumina microarrays and BeadStation instrument. Potential users should consult with the Facility (Kang Liu) prior to starting microarray experiments. For an excellent introduction to Illumina technology, see the narrated 10 slide tutorial from Illumina. In Illumina arrays, one uses Cy3-labeled cRNA to hybridize to bead-linked 50mer cDNA oligo probes arrayed on the slide surface. In overview, a user brings 100-500 ng of total RNA for each full genome analysis to the Facility, and receives the resulting data from replicate expression analysis in a tabular format ready for downstream analysis. Additionally, desired analysis such as ranked lists of genes found more than 2-fold different in the replicates is also provided.

Costs

The total cost (microarray plus labor) for analysis of each human RNA sample is $235, using a 24K gene array that represents the entire 23K RefSeq collection (nearly all of the nominally characterized human genes) and 1K of controls. A human 48K gene array that represents both the 23K RefSeq genes and also 23K other less well characterized transcribed sequences ($290 total analysis cost) is also available. Array analysis on similarly designed mouse gene arrays (24 or 48K) is the same cost as the human arrays. For full cost information, see Prices.

Arrays

HumanRef-8 v2 Expression BeadChip. This human gene array is also referred to as the 24K array, or the 8 x 1 array, as described below. The 24K array contains the entire 23K human RefSeq collection, as well as controls, and represents essentially all of the nominally characterized human genes. The large 24K Excel gene list is available from Illumina for download.

Human-6 v2 Expression BeadChip This array is also referred to as the 48K array or the 6 x 2 array, as described below. This array represents not only the entire 23K RefSeq collection present on the 24K array, but in addition, 23K other expressed sequences considered most likely to represent real genes, extracted from several databases. The large 48K Excel gene list is available for download.

MouseRef-8 Expression BeadChip . This mouse gene array represents a comparable set of genes to the HumanRef-8 array, and is also referred to as the mouse 24K array, or the 8 x 1 array.

Mouse-6 Expression BeadChip . This mouse gene array represents a comparable set of genes to the HumanRef-6 array and is also referred to as the mouse 48K array, or the 6 x 2 array.

RatRef-12 Expression BeadChip The RatRef-12 Expression BeadChip contains > 22,000 probes selected primarily from the NCBI RefSeq database It has 12 expression arrays on a single slide.

Practical considerations

High quality total RNA is required for array analysis. Excellent results have been obtained with either Trizol or Qiagen column isolated RNA. The 100-500ng of RNA (500 ng is optimal) used for labeled cRNA synthesis should be in no more than 9 microliters of water. Replicate array analysis (biological or technical) is essential, as even at the stringent level of P = 0.005, a single 48K data set could have over 200 random hits. It is far cheaper and faster to do replicate arrays and have sufficient power for significance analysis, than it is to follow up numerous false positives.

Each Illumina slide actually contains 8 separate full genomes (in 8 strips) for the 24K arrays, or 6 full genomes (in 6 pairs of strips) for the 48K arrays, and 12 full genomes for the rat array (see more). We use the term “genome” to avoid confusion caused by the term “array”, because most Illumina slides contain 6 or 8 genomes. Each genome (representing the 23K or 46K genes) is equivalent to what used to be called an array. Using a hybridization chamber with gaskets , each genome on the slide is separately hybridized to a different labeled cRNA preparation. Nonetheless, post-hyb processing takes place on the whole slide, so all of the genomes on the slide must be used in a single experiment. That is why it is stated on the price list that one must purchase six 48K genomes or eight 24K genomes (one full slide) at a time. In rare circumstances, it may be possible to find two users to share a slide. However, having 6 genomes on the 48K slide is very convenient, as this is just right for triplicate “A vs. B” comparison for this one-color analysis.

Generating labeled cRNA and array hybridization

Labeled cRNA is generated in the Facility from the whole RNA sample utilizing an RNA amplification kit developed for Illumina by Ambion. This follows the basic Eberwine technique, where first-strand cDNA is made after priming with a T7 promoter-oligo(dT) primer. Following second strand cDNA synthesis, T7-mediated transcription in the presence of biotin-16-UTP is used to generate biotinylated cRNA. Finally, this amplified, biotinylated cRNA is labeled by incubation with streptavidin-Cy3. A single RNA amplification/labeling reaction can generate sufficient probe for several array hybridizations. As described above, hybridization is carried out in special Illumina cassettes, with the standard hybridization and wash protocols recommended by the manufacturer.

Data analysis

Following the high-resolution scanning of the bar-coded array slides, the BeadStation software deconvolutes the data from an average of separate 30 beads per gene. A variety of control features are build into the arrays for normalization and quality control. The data are provided in a MIAME compatible tabular form. The Illumina BeadStudio software provides additional analysis capability, including statistical analysis of replicates, scatter plots with highlighting of user-definable differences in expression, and lists of altered genes with annotation.

For downstream data analysis, in addition to BeadStudio software, the Facility provides GeneSpring software, the gold standard for microarray analysis, capable of higher level analysis such as clustering and direct linkage to annotation. Training and assistance with array analysis software packages (typically not described as user-friendly) is available from the Informatics and Data Management Shared Resource. Additionally, for superimposing the expression results on biological pathways, and determining whether specific pathways or functions are being targeted, two additional (more user friendly) software packages linked to proprietary databases are available. These are 1) Ingenuity Pathways Analysis, for finding literature-based linkages between genes and placing them in common pathways, and 2) NextBio, which can compare overall changes seen in an array study with all other publicly available array results, finding novel similarities with other conditions, pathways, or cell types. Assistance with these and other analytical software (e.g., GSEA, Bioconductor, and currently evaluating GeneGo) for downstream array analysis are provided by Drs. Alexey Eroshkin and Roy Williams in the Informatics and Data Management Shared Resource.