mRNA-seq Whole Transcriptome Profiling of Fresh Frozen versus Archived Fixed Tissues

Background: The main bottleneck for genomic studies of tumors is the limited availability of fresh frozen (FF) samples collected from patients, coupled with comprehensive long-term clinical follow-up. This shortage could be alleviated by using existing large archives of routinely obtained and stored Formalin-Fixed Paraffin-Embedded (FFPE) tissues. However, since these samples are partially degraded, their RNA sequencing is technically challenging. Results: In an effort to establish a reliable and practical procedure, we compared three protocols for RNA sequencing using pairs of FF and FFPE samples, both taken from the same breast tumor. In contrast to previous studies, we compared the expression profiles obtained from the two matched sample types, using the same protocol for both. Three protocols were tested on low initial amounts of RNA, as little as 100 ng, to represent the possibly limited availability of clinical samples. For two of the three protocols tested, poly(A) selection (mRNA-seq) and ribosomal-depletion, the total gene expression profiles of matched FF and FFPE pairs were highly correlated. For both protocols, differential gene expression between two FFPE samples was in agreement with their matched FF samples. Notably, although expression levels of FFPE samples by mRNA-seq were mainly represented by the 3'-end of the transcript, they yielded very similar results to those obtained by ribosomal-depletion protocol, which produces uniform coverage across the transcript. Further, focusing on clinically relevant genes, we showed that the high correlation between expression levels persists at higher resolutions. Conclusions: Using the poly(A) protocol for FFPE exhibited, unexpectedly, similar efficiency to the ribosomal-depletion protocol, with the latter requiring much higher (2-3 fold) sequencing depth to compensate for the relative low fraction of reads mapped to the transcriptome. The results indicate that standard poly(A)-based RNA sequencing of archived FFPE samples is a reliable and cost-effective alternative for measuring mRNA-seq on FF samples. Expression profiling of FFPE samples by mRNA-seq can facilitate much needed extensive retrospective clinical genomic studies. Overall design: We perform an unbiased evaluation of RNA-seq of archived tumor tissues by comparing the same library preparation methods for both FF and FFPE matched tumor samples and for small amounts of total RNA starting material. We have 3 matched FF/FFPE tumor samples with a moderate archival time of about 4-5 years (T1=T3), and additional 3 FFPE tumor samples archived for more than 10 years (T4-T6). all samples were tested with two protocols: illumina Truseq RNA after poly(A) selection (mRNA-seq); and Truseq after ribosomal depletion (RiboZero). Several initial amounts of starting material was tested for eacg protocol.

190

Bossel Ben-Moshe N, Gilad S, Perry G, Benjamin S, Balint-Lahat N, Pavlovsky A, Halperin S, Markus B, Yosepovich A, Barshack I, Gal-Yam EN, Domany E, Kaufman B, Dadiani M