eLabNotebook > Nucleic Acid Prep & Purification > Target Preparation > GeneChip Expression
Analysis on Biomek® 3000 ArrayPlex

Target Preparation for GeneChip* Expression Analysis Using the Biomek® 3000 Laboratory Automation Workstation and ArrayPLEX Application

Zhu Zhu, Dan Allison, Yu Suen, Handy Yowanto, Tim Lancaster, Jeremy Smith and Keith Roby,
Beckman Coulter, Inc., Fullerton, California.

Abstract

Microarray technology has become a standard tool for gene expression researchers, enabling the simultaneous assessment of changes in gene expression across the whole genome. However, the process of target preparation for the GeneChip expression analysis is time consuming, labor intensive, and prone to human errors. We have developed a solution, the ArrayPlex application, to automate the target preparation process on the Biomek 3000 Laboratory Automation Workstation (Beckman Coulter). The ArrayPLEX application is part of the GenomeLab family of applications (Beckman Coulter).

Figure 1: B3k with the ArrayPlex Application

The highly automated process comprises three methods: cDNA Synthesis, in vitro Transcription (IVT), and Normalization and Fragmentation. It allows for processing a partial or a full 96-well plate of samples from 1 – 2 µg of total RNA to the biotin-labeled complementary RNA (cRNA) using Affymetrix’ GeneChip Expression 3’-Amplification reagents and the Agencourt® RNAClean™ kit.

The yield, of the target cRNA generated using these methods, was measured by OD260 with a DTX 880 Multimode Detector (Beckman Coulter). The quality of the cRNA was analyzed on the Agilent 2100 Bioanalyzer. The cRNA samples were normalized to a target concentration of 0.625 µg/µl, fragmented, and hybridized to Affymetrix’ Human Genome U133 Plus 2.0 Arrays.

Figure 1. Diagram of the B3K deck setup.

Figure 2 B3K Deck configurations for the ArrayPLEX Application with a Thermocycler
Note: The deck configuration for the ArrayPLEX Application without a Thermocycler is not shown. The only difference to the deck shown above is that the regular Biomek 3000 position (P5) will be at PTC1.

B3K Method                 Work Flow               Time

Number of Samples to Process

The highly automated methods process a full or partial 96-well plate of samples in multiples of 8 in a single run. The prompt at the beginning of each method allows the users to enter the number of columns to process (from 1 – 12).

Figure 3: Entering the number of columns to process at the beginning of Biomek 3000 ArrayPLEX methods.

Steps in Method

Each of the three methods comprises two independent steps (highlighted below in the red frames): cDNA Synthesis/cDNA Cleanup, cRNA Synthesis/cRNA Cleanup and Quantitation/Normalization and Fragmentation. Users can save samples after cDNA Synthesis, cRNA Synthesis or Quantitation and run the remaining processing steps at a later time. The decks need to be refreshed before cDNA and cRNA cleanup.

Figure 4: cDNA Synthesis/cDNA Cleanup, cRNA Synthesis/cRNA Cleanup and Quantitation/Normalization and Fragmentation steps

Figure 5:  Deck Layout for cDNA Synthesis Method

Figure 6: Deck Layout for IVT method

Figure 7: Deck Layout for Fragmentation Method

The concentration of cRNA was calculated based on OD260 measured using an offline DTX 880 Multimode Detector. A calculation template specifically created for the ArrayPlex application was used to correct the reader data with the dilution factor and adjust the concentration unit to µg/µl. The corrected cRNA concentration was then exported to the B3K for normalization. The normalization target was 0.625µg/µl with a volume of 35 µl. The concentration of the normalized samples was verified to ensure that it met the target concentration prior to fragmentation.

Validation of the Methods

Validation I

• The objective was to achieve Affymetrix Premier Application provider status.
• A total of eight partial plates were run on two workstations with a thermocycler.
• Each plate had 24 samples, 12 of 1 µg and 12 of 2 µg human Hela total RNA, randomly distributed in columns 1 – 3.
• Six of the cRNA samples from each plate, three from 1 µg and three from 2 µg total RNA, were randomly selected to hybridize to the Human Genome U133 Plus 2.0 arrays.

Validation II

• The objective was to demonstrate that the methods could process a full 96-well plate.
• A total of two full plates were processed, one on a workstation with a thermocycler and one on a workstation with an offline thermocycler.
• Each plate had 24 samples, 12 of 1 µg and 12 of 2 µg human Hela total RNA, randomly distributed in columns 1, 6 and 12. Water was used to fill out the remaining wells in the cDNA Synthesis and IVT methods, while caffeine solutions with appropriate OD260 readings were used in the Fragmentation method.

cRNA Yield and CV

Table 1 The cRNA yield was calculated based on a 40 µl volume. The normalization target was 0.625 µg/µl. OD260 readings of the caffeine solutions were also used when calculating the P. Norm Avg. and CV for the full plates. The tables show consistent cRNA yield, post normalization average and CV among the plates. The cRNA yields were comparable across the columns of the full plates. Plate 1.1, 2.1, 3.1 and 4.1: plates ran on workstation 1. Plate 1.2, 2.2, 3.2 and 4.2: plates ran on workstation 2. TC: the workstation with a thermocycler. No TC: the workstation with an offline thermocycler.

Figure 8 The average cRNA yield was 83.1 µg from 1 µg of total RNA (shown in blue) and 135.6 µg from 2 µg of total RNA (shown in red) for the 8 plates run in Validation I. All samples meet the post normalization target (21.9 µg) and the average concentration was 0.69 µg/µl. The average post normalization CV was 4.6%. The error bars show the highest and lowest yields in each group.

Figure 9.  cRNA samples were analyzed on the Agilent 2100 Bioanalyzer. The samples were diluted 1:4 to fit in the range of the Agilent RNA 6000 Nano Assay. The Agilent RNA 6000 Nano Ladder was used as the marker. The peak sizes are 0.2, 0.5, 1.0, 2.0, 4.0 and 6.0 kb for peaks 1 through 6. The electropherogram shown is a representative of the samples. Similar non-degraded cRNA profiles were observed with samples amplified from 1 µg and 2 µg of total RNA. The average peak size was approximately 1580 nt.

GeneChip Analysis

Percent Present

Percent False Change (log2)

Table 2 A total of 48 cRNA samples of Validation I, 3 of 1 µg and 3 of 2 µg Hela total RNA from each plate, were hybridized to the Affymetrix Human Genome U133 Plus 2.0 Array. The table lists the Percent Present and Percent False Change generated only from U133 subset A. The data showed very good Percent Present Call and reproducibility.

GeneChip Analysis- Consistent 3’/5’ Signal Ratio

Figure 10: The line graph show consistent 3’/5’ Signal Ratio across all the plates with an average of 1.07 for GAPDH and 2.11 for Beta-Actin.

GeneChip Analysis: Array Signal Correlation

Figure 11: The blue diagonal bars indicate a two-fold change. The scatter plots show equivalent performance of cRNA samples prepared from 1 or 2 µg total RNA in different runs on same or different workstation.
A: 2 µg vs. 2 µg, same workstation.
B: 1 µg vs. 1 µg, different workstation.
C: 1 µg vs. 2 µg, same workstation.
D: 1 µg vs. 2 µg, different workstation.

Validation I Results vs. Specifications Set by Affymetrix

eLabNotebook Sitemap

©1998 - 2006 Beckman Coulter, Inc.