Cancer Diagnosis Program | Biorepositories and Biospecimen Research Branch

Search BRD Papers

RNA Sequencing of Formalin-Fixed, Paraffin-Embedded Specimens for Gene Expression Quantification and Data Mining.

Author(s): Guo Y, Wu J, Zhao S, Ye F, Su Y, Clark T, Sheng Q, Lehmann B, Shu XO, Cai Q

Publication: Int J Genomics, 2016, Vol. 2016, Page 9837310

PubMed ID: 27774452 PubMed Review Paper? No

Purpose of Paper

Conclusion of Paper

Studies

1. Study Purpose

   This study examined the effects of ribosomal RNA (rRNA) depletion using two different methods on RNAseq data and the correlation in expression with NanoString results in four FFPE triple-negative breast cancer (TNBC) specimens. FFPE sections of the four specimens were stored in a vacuum at 4˚C for 8-9 years before RNA extraction. Details of formalin-fixation and paraffin-embedding were not provided. Total RNA was extracted using the miRNeasy FFPE kit. RNA was quantified by spectrophotometer and integrity was determined by Bioanalyzer. Specimens were rRNA-depleted in duplicate with Ribo-Zero or RNase H before library construction using TruSeq RNAsample Prep kit v2. Libraries were sequenced by BGI Americas. The authors state NanoString nCounter data was obtained for 302 genes using the same specimens, but do not state any details of how this was obtained.

   Summary of Findings:

   Specimens that were rRNA depleted using RNAse H rather than RiboZero produced a higher average number of reads (21.4 versus 16.1 million per sample), had a higher percentage of reads that aligned to the coding region (71.6% versus 34.7%), had a higher mapping quality in the coding region (45 versus 29), and a lower percentage of reads mapped to rRNA (P=0.03), but a lower mapping quality in the noncoding region (34 versus 47). The average GC content was closer to the expectation of 50% when depleted using RNAse H rather than Ribo-Zero (50.6% versus 72.3%). One specimen depleted with RNAse H was considered to be an outlier as it had the lowest GC content (39.8%) and coding region alignment rate (42.6%). Cluster analysis clustered specimens first by patient and then by rRNA depletion method with one RNAse H-depleted specimen outlier. The TNBC subtype was more consistently assigned in RNAse H-depleted specimens than RiboZero-depleted specimens (3/4 versus 2/4), with the outlier producing the nonmatching subtype for one RNAse H-depleted pair. However, the average correlation in results with NanoString was higher for RiboZero-depleted than RNAse H-depleted specimens (0.59 versus 0.49). Although more RNAs were detected at a threshold of >0 or >2 when depleted with RiboZero, RNAse H-depleted specimens had more RNA detected at higher detection thresholds. More SNVs were detected in RNAse H-depleted than RiboZero-depleted specimens (57,747 versus 525), but using both methods the ratio of transitions to transversion mutations (Ti/Tv) was high (3.74 for RNAse H and 2.84 for RiboZero) and the consistency rates were less than 40%, indicating the SNV results are unreliable.

   Biospecimens

   • Tissue - Breast

   Preservative Types

   • Formalin

   Diagnoses:

   • Neoplastic - Carcinoma

   Platform:

   Analyte	Technology Platform
   RNA	DNA microarray
   RNA	Next generation sequencing

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Next generation sequencing Specific	Technology platform	NanoString
   Next generation sequencing Specific	Template modification	Ribo-Zero depleted RNAse H depleted

   View more