RNA-Seq Accuracy Comprehensive Assessment, Duplicability and Knowledge Content by the Sequencing Internal Control Consortium

Abstract

We gift primary results from the Sequencing Internal Control (SEQC) project, coordinated by the America Food and Drug Administration. Examining Illumina HiSeq, Life Technologies SOLiD and Roche 454 platforms at multiple laboratory sites victimization reference RNA samples with intrinsical controls, we have a tendency to assess RNA sequencing (RNA-seq) performance for junction discovery and differential expression identification and compare it to microarray and quantitative PCR (qPCR) information victimization complementary metrics. in the slightest degree sequencing depths, we have a tendency to discover unannotated exon-exon junctions, with >80% valid by qPCR. We discover that measurements of relative expression area unit correct and reproducible across sites and platforms if specific filters area unit used. In distinction, RNA-seq and microarrays don't offer correct absolute measurements, and gene-specific biases area unit determined for all examined platforms, together with qPCR. Activity performance depends on the platform and information analysis pipeline, and variation is giant for transcript-level identification. The whole SEQC information sets, comprising >100 billion reads (10Tb), offer distinctive resources for evaluating RNA-seq analyses for clinical and regulative settings.