Overview
This article describes a TruSeq (Illumina, Inc. San Diego, CA) RNA library preparation method for FFPE samples that uses the Biomek automation platform (Beckman Coulter Life Sciences, Indianapolis, IN) in which 96 libraries are constructed in seven hours. The Biomek platform provides a solution for high throughput library construction, producing sensitive and reproducible sequencing data, which facilitates biomarker discovery in archival FFPE tissues.
Introduction
Next Generation Sequencing (NGS) technology facilitates high throughput, high speed and cost-effective sequencing of DNA and/or RNA. RNA library construction from archival FFPE tissue specimens can be automated utilizing the Beckman Coulter Biomek FXP Liquid Handler. RNA is extracted using the Beckman Coulter Agencourt FormaPure kit and the library preparation method is based on the Illumina TruSeqRNA Sample preparation protocol.
The Biomek/TruSeq RNA method is comprised of three parts:
1. mRNA purification, fragmentation and cDNA synthesis
2. cDNA library construction (end repair, A-tailing and adaptor ligation)
3. PCR amplification and product purification
The data presented here benchmarks an automated approach that uses the Biomek FXP Liquid Handler against a standard manual preparation. Archival FFPE tissue specimens have lower RNA yield and are highly degraded, which represents a major challenge in developing robust methodologies for sequencing. The findings of the study highlighted that preparation on the Biomek automated platform provided researchers with a more reliable, more streamlined approach for generating libraries from precious clinical samples with a faster turnaround time.
Methods and Results
RNA Preparation
Three tumour/normal pairs of archival FFPE prostate tissue specimens were selected and scored. Total RNA was extracted according to the protocol of the Agencourt FormaPure kits (Beckman Coulter Life Scienc¬es). Each sample was eluted in 55 μL dH2O. RNA quality was determined on the Bioanalyzer 2100 system (Agilent Technologies Inc., Santa Clara, CA). Figure 1 shows the typical degraded profile of RNA isolated from archival FFPE tissue. The RN number was ≤ 2.4, which indicates the high level of degradation associated with these tissues.
TruSeq cDNA Library Preparation
Fragmented RNA was reverse-transcribed to make the first strand cDNA. The first strand cDNA was then used as a template to generate double stranded DNA. The dsDNA template was blunt-ended and adenlyated at the 3’ ends to enforce strict ligation of the adapters. The adapter modified templates were subjected to 15 rounds of amplification to enrich and selectively amplify DNA containing adapter molecules. After the final clean up step, the purified DNA samples were eluted in 30 μL of Resuspension Buffer. 1 μL of the library was used to assess the quality of the library generated using Agilent High Sensitivity DNA kit on the Bioanalyzer 2100 system (Agilent Technologies).
cDNA Library Quality Cleanup
Initial quality review showed an average peak size for the amplified cDNA library at approximately 260 bp, however, there was a 120 bp non-specific adaptor amplified peak also present in all 12 samples. This represents adaptor-dimer contamination, which can lead to junk reads when sequencing. In some cases, an additional fragment greater than 400 bp was also detected, which could represent single-stranded library products that have self-annealed.
In order to improve library quality, a two-step size selection clean-up protocol was implemented. In the first clean-up step, a 0.7X ratio of AMPure XP bead solution (Beckman Coulter Life Sciences) was added to each sample. Under this condition, only larger fragments greater than 500 bp DNA were bound to the beads. The bead-bound DNA was discarded and the supernatant treated with a 1.1X volume of AMPure beads to rebind DNA fragments between 150 bp – 400 bp. Figure 2 shows the overlaid QC plots of the pre- and post-clean up libraries for tumor/benign pair A with a side-by-side comparison of the Biomek and manual libraries.
Single End Sequencing Performance
Sequencing was performed on the Illumina HiSeq platform. 50 bp single reads were mapped using the TopHat program (Institute of Genetic Medicine, Johns Hopkins University Center for Bioinformatics and Computational Biology (CBCB), Baltimore, Md.) and transcript abundance in FPKM units (fragments per kilobase of mRNA per 106 reads) were calculated using Cufflinks (CBCB). Multiplexed sequencing was performed such that each lane contained three samples. RNASeq library construction/sequencing was successful in 12/12 samples with aligned reads ranging from 21 per cent to 72 per cent relative to the total number of reads.
Reproducibility was measured by calculating the Pearson Correlation from transcript abundance values, presented as Log2 [FPKM] values. TruSeq libraries prepped both by the Biomek automated platform and manually on the bench, exhibited good concordance, as evidenced in the scatter plots depicted in Figure 3.
The total number of transcripts is also displayed in Figure 3. In particular, the poor library quality of tumor/normal pair C is verified in the corresponding Pearson correlation coefficients.
A wide ranging panel of endogenous controls was found to show good concordance between the manual and Biomek sequenced libraries, with a Pearson coefficient of 0.99 calculated. The insets are plots of the average Log2 [FPKM] values of all samples for each library preparation method, which highlights the stability of these endogenous controls across all sample types, irrespective of preparatory method or tissue morphology.
Conclusion
This study shows that RNASeq library preparation carried out on the automated high throughput Biomek platform, results in sensitive and re¬producible sequencing data, which facilitates biomarker discovery in archival FFPE tissue. Going forward it is envisaged that new coding and non-coding transcripts, as well as gene signaling networks that strongly associate with prostate cancer progression will be identified.
The Biomek automation platform, capable of constructing 96 libraries in seven hours, offers a viable high throughput alternative to traditional manual bench preparation of RNA libraries for sequencing. The Biomek FXP automation platform provides a high-throughput workflow and a more reliable way to generate libraries from precious clinical samples with faster turnaround time. In addition, the same method can be used with fresh tissue or other sample types for library construction.
Reference
1. DeJonge, HJM et al., Evidence Based Selection og Housekeeping Genes, PLoS One, 2007, 2(8), e898. Beckman Coulter, the stylized logo, Biomek, Agencourt, AmPure XP, and FormaPure, are registered trademarks of Beckman Coulter and are registered in the USPTO. The Biomek FXP is registered as a medical device in S. Korea, and is not intended for diagnostic applications in other countries. All other trademarks are the property of their respective owners.
Leave a Reply