Applied Biosystems Expression Array System

The ABI system is a relatively new entry to the market. It differs from other single-color oligo arrays in several ways:

• Probes are designed against the Celera Genomics genomes, as opposed to the public genomes utilized by other manufacturers.
• Probes are single 60-mer oligos
• Printed spots have a separate flourescent dye attached to them, to ensure proper grid alignment
• Samples are tagged with a chemiluminescent dye, which is then detected as signal
• For more information, see the System Profile (PDF)

AB currently offers three microarrays - Human, Mouse, and Rat. The WCIMC stocks Human and Mouse arrays, but investigators who wish to use Rat arrays will need to purchase an entire pack of four.

Detailed information about the design of these microarrays can be found here: AB1700-array-design.pdf

Detailed information about the AB1700 system and how it works can be found here: 1700-system.pdf

AB Software

Applied Biosystems intentionally designed their system to not require any particular software - data is exported in a flexible format that can be read by nearly any statistical or microarray package.

The PANTHER database contains sequence information, TaqMan assays, and the most current annotations for every probe on an AB microarray.

AB Data

Data is exported from the 1700 scanner as a tab-delimited text file which can be easily loaded into your favorite analysis package (or Excel). For a detailed description of this data, see the VMSR ABI Analysis Manual, located on your ABI project page. Briefly, the most important columns are:

• Signal: The background-corrected, unnormalized signal of the feature
• S/N: The Signal-to-Noise ratio expresses the confidence of the feature's 'detectibility'. A S/N of 3 represents a 3 SDEV confidence (99.9%) that the measurement is real.
• Flags: Correspond to the errors in measurement encountered while scanning. Larger numbers mean more serious errors - anything less than 8000 can be safely ignored.

As part of the standard analysis performed with hybridization, the VMSR will perform a quantile normalization of your data (see Data Normalization). This normalized data will be named something similar to "QuantileNormalizedExpression.csv", and can be read directly into Excel. Note that this data file has been log2 transformed!

If your data set includes replicates, we will also perform a t-test between each group of replicates (see Significance Analysis). The filename of your t-test results will start with "DetectPct50FCpvalFDR", and one file will be generated for each t-test run. This file will include the fold change, p-value, and Benjamini-Hochberg corrected p-value for each feature analyzed.