Science Inventory

The effects of age-in-block on RNA-seq analysis of archival formalin-fixed paraffin-embedded (FFPE) samples


Wehmas, L., S. Hester, V. Bhat, B. Chorley, G. Carswell, AND C. Wood. The effects of age-in-block on RNA-seq analysis of archival formalin-fixed paraffin-embedded (FFPE) samples. NIEHS Genomics Day, RTP, NC, May 12, 2016.


Biorepositories have immense potential value in characterizing molecular targets of environmental chemicals. The goal of this study was to evaluate dose-response metrics for RNA-seq data generated from recently archived and older formalin-fixed paraffin-embedded (FFPE) samples.


Archival samples represent a vast resource for identification of chemical and pharmaceutical targets. Previous use of formalin-fixed paraffin-embedded (FFPE) samples has been limited due to changes in RNA introduced by fixation and embedding procedures. Recent advances in RNA-sequencing (RNA-seq) technology offer a way to overcome at least some of the issues associated with older degraded FFPE RNA. In this experiment, we examined the quality of transcriptomic dose-response data from FFPE RNA samples compared with matched frozen (FROZ) samples from two experiments < 2 years old (Study 1) and >20 years old (Study 2). Experimental treatments were di(2-ethylhexyl)phthalate (DEHP) and dichloroacetic acid (DCA) for Studies 1 and 2, respectively. Samples were sequenced following total RNA collection and ribo-depletion on an Illumina Hi-seq 2500 sequencer. The total gene counts (summed across 31,252 possible gene targets and averaged to DEHP or DCA treatment groups) from FFPE samples were reduced by 35% in Study 1 and 97% in Study 2 compared to FROZ samples. High concordance in quality metrics and dose responses were observed for Study 1, including the magnitude and number of differentially expressed genes (DEGs) between FFPE and FROZ samples (r2 = 0.99). Pathway enrichment analysis (IPA) for Study 1 identified 90% overlap of FFPE with frozen samples, and little difference (2%) in FFPE benchmark dose estimates (BMDExpress) for preselected target genes versus FROZ. In contrast, poor agreement in global DEGs, enriched pathway analysis, and bench mark dose estimates were observed for several preselected target genes between FFPE and FROZ samples for Study 2. However, FFPE read counts for more highly abundant gene transcripts like Cyp4a10 and Cyp4a14 were still positively correlated with FROZ samples (r2 = 0.84). These findings highlight potential applications and challenges in using RNA-seq data from FFPE samples and emphasize the need for improved methodology in obtaining data from much older, poorer quality FFPE samples. This work will assist in advancing the use of archival resources in chemical safety. Disclaimer: The research described in this article has been reviewed by the National Health and Environmental Effects Research Laboratory, US Environmental Protection Agency, and approved for publication. Approval does not signify that the contents necessarily reflect the views and the policies of the Agency.

Record Details:

Product Published Date: 05/12/2016
Record Last Revised: 05/13/2016
OMB Category: Other
Record ID: 314270