You are here:
Evaluating Quality of Aged Archival Formalin-Fixed Paraffin-Embedded Samples for RNA-Sequencing
Citation:
Wehmas, L., S. Hester, V. Bhat, B. Chorley, G. Carswell, AND C. Wood. Evaluating Quality of Aged Archival Formalin-Fixed Paraffin-Embedded Samples for RNA-Sequencing. Environmental Mutagenesis and Genomics Society, Kansas City, MO, September 24 - 28, 2016.
Impact/Purpose:
Archival formalin-fixed paraffin-embedded (FFPE) samples offer a vast, untapped source of genomic data for biomarker discovery. However, the quality of FFPE samples is often highly variable, and conventional methods to assess RNA quality for RNA-sequencing are not informative. Our goal was to investigate the relationship between RNA fragment size distribution, RNA sequencing quality parameters, and differentially expressed gene profiles to identify a better method to assess RNA quality prior to RNA-sequencing.
Description:
Archival formalin-fixed paraffin-embedded (FFPE) samples offer a vast, untapped source of genomic data for biomarker discovery. However, the quality of FFPE samples is often highly variable, and conventional methods to assess RNA quality for RNA-sequencing (RNA-seq) are not informative. In this study, we evaluated transcriptomic dose-response data from 40 pairs of FFPE and frozen (FROZ) RNA from <2 year-old, “high-quality” (Study 1) and >20 year-old, “low-quality” (Study 2) samples. Our goal was to investigate the relationship between RNA fragment size distribution, RNA-seq quality parameters, and differentially expressed gene (DEG) profiles. Agilent Bioanalyzer trace analysis was used to evaluate RNA quality prior to RNA-seq by quantifying the percentage of RNA fragments greater than 100, 150, and 200 nucleotides (DV100, DV150, and DV200 respectively). For RNA-seq analysis, samples were ribo-depleted and sequenced on the Illumina Hi-seq 2500 platform. High concordance in RNA-seq quality metrics, DEGs, and dose responses were observed between FFPE and FROZ samples in Study 1 but not Study 2 despite similar mean RNA integrity numbers (Study 1: 2.2±0.28; Study 2: 2.4±0.50). Spearman correlation analysis of data combined from both studies revealed that DV100 values correlated best with several RNA-seq metrics (ρ=0.72−0.95). Samples with a DV100 <81% (20/24 Study 2 FFPE samples) showed insufficient and inconsistent global levels of gene detection (<7000) for RNA-seq analysis. These findings highlight potential applications and challenges in using RNA-seq data from FFPE samples. This abstract may not necessarily reflect official Agency policy.