You are here:
Use of fragment analysis to predict sequencing success for low-quality RNA
Citation:
Wehmas, L., S. Hester, AND C. Wood. Use of fragment analysis to predict sequencing success for low-quality RNA. Society of Toxicology, Baltimore, MD, March 12 - 16, 2017.
Impact/Purpose:
RNA obtained from archival tissue samples can vary widely in quality, limiting their use in genomic research. Standard methods for measuring RNA integrity do not adequately assess aged formalin-fixed paraffin-embedded (FFPE) sample RNA suitability for RNA sequencing. We examined whether the DV100-200 metric, which quantifies the percent of RNA fragments greater than 100-200 nucleotides, better characterizes FFPE RNA quality for sequencing compared to RNA integrity number (RIN). The results identify that RNA fragmentation analysis is more informative than RIN in characterizing FFPE RNA quality and will help establish a metric for prioritizing FFPE samples for RNA sequencing and genomic analyses. This research supports methods and tools for AOP discovery, which addresses goals of Task 1.1g -- AOP Descriptions for Cancer.
Description:
RNA obtained from archival tissue samples can vary widely in quality, limiting their use in genomic research. Standard methods for measuring RNA integrity do not adequately assess aged formalin-fixed paraffin-embedded (FFPE) sample RNA suitability for RNA sequencing. Here, we examined whether the DV100-200 metric, which quantifies the percent of RNA fragments greater than 100-200 nucleotides, better characterizes FFPE RNA quality for sequencing compared to RNA integrity number (RIN). We evaluated the DV100-200 and RIN for FFPE mouse liver samples of varying age in paraffin block (1-21 years) and length of time in formalin (0, 18 hours, and 3 months) (total n=76) which were compared to several RNA sequencing quality indicators (read quality, mapping, gene detection). DV analysis and RIN were calculated using an Agilent 2100 Bioanalyzer prior to sequencing on an Illumina Hi-seq 2500 instrument. RNA sequencing analysis was performed using Partek Flow (v5). Spearman correlation analysis revealed that DV200 associated best with most if not all RNA sequencing quality parameters compared to DV100, DV150, and RIN across all FFPE samples analyzed. Notably, DV200 correlated far better than RIN for global gene detection (ρ= 0.82 and ρ= 0.15, respectively). The RIN averaged 2.4 ± 0.8 across all samples, yet percent genome coverage ranged from 0.1%-14% and genes detected ranged from 0.3-16k. Using the current DV200>30% threshold recommended by Illumina, would have excluded 15% of our samples with good gene detection (~12.5k). DV100>80% performed better than DV200>30% in distinguishing 2 & 21 yr FFPE samples suitability for sequencing well but not samples with differing time in formalin. These findings indicate that RNA fragmentation analysis is more informative than RIN in characterizing FFPE RNA quality for sequencing but also highlight the need for additional criteria or metrics. These results will aid in advancing the use of archival FFPE samples for genomic analyses. This abstract does not reflect official Agency policy.