Cancer Diagnosis Program | Biorepositories and Biospecimen Research Branch

Search BRD Papers

Accurate RNA Sequencing From Formalin-Fixed Cancer Tissue To Represent High-Quality Transcriptome From Frozen Tissue.

Author(s): Li J, Fu C, Speed TP, Wang W, Symmans WF

Publication: JCO Precis Oncol, 2018, Vol. 2018, Page

PubMed ID: 29862382 PubMed Review Paper? No

Purpose of Paper

Conclusion of Paper

Studies

1. Study Purpose

   This study compared the uniformity of coverage and the accuracy of sequencing data among different library preparation methods for RNA sequencing (including total RNA sequencing after rRNA depletion and exon capture-based sequencing with and without demodification).  The results generated from RNA isolated from FFPE breast tumor specimens using each method were compared to those from case- matched frozen specimens.  Nine breast tumor specimens were diced into 1-2 mm segments and either placed in RNAlater and stored at -80°C (frozen), or fixed in 10% neutral buffered formalin and paraffin-embedded (FFPE; details not provided). Breast cancer cases included five hormone receptor (HR)-positive/HER2-negative, one HR-positive/HER2-positive, and three triple negative cases. RNA was extracted from the frozen specimen using the RNeasy Mini Kit, and from 10 µm thick sections of the FFPE specimen using the HighPure FFPE RNA Isolation Kit. RNA sequencing libraries were constructed from RNA from frozen specimens using a poly-A mRNA enrichment method and the TruSeq RNA Sample Prep Kit v2. Total RNA libraries were also constructed from RNA extracted from case-matched frozen and FFPE specimens after (i) rRNA depletion with the Ribo-Zero Magnetic Gold kit using the Truseq Stranded Total RNA Sample Prep Kit, (ii) after depletion of rRNA with RNAse H using the KAPA Stranded RNA-Seq Kit with RiboErase, and (iii) after whole transcriptome amplification with Sensation Plus Amplification kit, rRNA depletion with Ribo-Zero Magnetic Gold Kit and library construction with the TruSeq RNA Sample Prep Kit v2. Additionally, exon capture libraries were prepared from fragmented RNA, demodified RNA and demodified RNA amplified using random primers with the Truseq Access RNASeq Kit. All libraries were pair-end sequenced on a HiSeq 2000 instrument.

   Summary of Findings:

   As expected, RNA extracted from FFPE specimens was highly degraded relative to frozen specimens, with RINs of 1.2-2.2 versus 6.7-9.3, respectively. Nevertheless >49 million reads were generated for each FFPE and frozen specimen.  RNA extracted from FFPE specimens generated fewer exonic and more intronic reads than case-matched frozen specimens (P<0.001, both), although the number of reads that mapped to intragenic regions was comparable among FFPE and frozen specimens.  Compared to the other library preparation protocols evaluated, library preparation  with TruSeq RNA Sample Prep Kit v2 after transcriptome amplification with the Sensation Plus Amplification kit and rRNA depletion with Ribo-Zero Magnetic Gold Kit (method iii) resulted in lower concordant pair alignment than with methods that did not include whole transcriptome amplification (P<0.001).  Uniformity of coverage was highest (with the lowest coefficient of variance) when FFPE libraries were generated from the Ribo-Zero Magnetic Gold Kit rRNA depleted RNA with TruSeq RNA Sample Prep Kit v2 (method i) or from RNAse H rRNA depleted RNA with KAPA Stranded RNA-Seq Kit (method ii) and the least uniform coverage occurred when the exon capture method was used. For the exon capture method, 26.5% of variance was attributable to the use of exon capture probes, and 20.6% was attributable to a combination of FFPE processing and biologic differences. Nevertheless, hierarchical clustering properly grouped most FFPE and frozen specimens based on hormone receptor status. When gene expression was compared using total RNA libraries generated with RNA isolated from FFPE and frozen specimens, log values centered around zero, but when the Total RNA libraries were compared with the exon capture libraries the log values were not centered around 0 for low expression transcripts. Importantly, the highest rank correlation between frozen libraries (multiple methods) and FFPE specimens was achieved using the KAPA Stranded RNA-Seq Kit with RiboErase.  Demodification of the RNA before generation of exon capture RNA libraries increased correlations with total RNA libraries but the difference was not significant. When compared to frozen miRNA libraries, the FFPE RNA using KAPA Stranded RNA-Seq Kit with RiboErase resulted in the fewest false positives for differential expression and the lowest dissimilarity score; while exon capture libraries resulted in the largest number of false positives. Finally, Total RNA libraries constructed using the KAPA Stranded RNA-Seq Kit with RiboErase provided the best discrimination between tumor types.

   Biospecimens

   • Tissue - Breast

   Preservative Types

   • Frozen
   • RNAlater
   • Formalin

   Diagnoses:

   • Neoplastic - Carcinoma

   Platform:

   Analyte	Technology Platform
   RNA	Automated electrophoresis/Bioanalyzer
   RNA	Next generation sequencing

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Biospecimen Preservation	Type of fixation/preservation	Formalin (buffered) RNAlater
   Next generation sequencing Specific	Nucleic acid amplification	Poly-A mRNA enrichment method and the TruSeq RNA Sample Prep Kit v2 Total RNA libraries constructed after rRNA depletion with the Ribo-Zero Magnetic Gold kit using the Truseq Stranded Total RNA Sample Prep Kit Total RNA library constructed after depletion of rRNA with RNAse H using the KAPA Stranded RNA-Seq Kit with RiboErase Sequencing library constructed after whole transcriptome amplification with the Sensation Plus Amplification kit, rRNA depletion with the Ribo-Zero Magnetic Gold Kit and TruSeq RNA Sample Prep Kit v2 Exon capture library generated from fragmented RNA using the Truseq Access RNASeq Kit Exon capture library generated from demodified RNA using the Truseq Access RNASeq Kit Exon Capture library generated from demodified RNA amplified using random primers with the Truseq Access RNASeq Kit

   View more