Simple and robust NGS RNA-based assay to assess impact of VUS on splicing. E. Girard1, J. Tarabeux2, E. Bernard1, A. Collet3, A. Legrand3, V. Moncoutier3, C. Dehainault3, J. P. Vert1, D. Stoppa-Lyonnet2,4, N. Servant1, C. Houdayer2,4 1) Institut Curie, INSERM U900, Mines ParisTech, Paris, France; 2) Institut Curie, Service de Génétique, INSERM U830, Paris, France; 3) Institut Curie, Service de Génétique, Paris, France; 4) Université Paris Descartes, Sorbonne Paris Cité, France.

   One of the key issues raised in molecular diagnosis is the correct interpretation of variants of unknown significance (VUS). Each VUS can potentially affect pre-mRNA splicing and be deleterious via disruption / creation of consensus sequences. Assessing the impact of VUSs on splicing is a central issue in order to determine their pathogenicity. Complete splicing defects can be easily demonstrated e.g. for exonic variants and for intronic variants when an exonic SNP is used as an indirect marker of instability. Conversely, defects affecting a fraction of the transcripts produced are more difficult to interpret as quantitative evaluation of partial splicing defects cannot be based on inspection of Sanger traces. Suitable quantitative methods such as pyrosequencing are available but not used in routine diagnosis. We therefore developed a targeted-single gene RNAseq strategy which lends itself as a simple and robust method providing qualitative as well as quantitative information on VUS impact at the RNA level. Patients with distinct RB1 splice mutations previously ascertained by Sanger sequencing were tested (resulting in out-of-frame/in-frame exon skipping and/or intronic retention). The complete cDNA sequence was amplified in a single PCR experiment and sequenced on a PGM (Life Technologies) following enzymatic shearing, barcoding and adaptor ligation. Reads were trimmed to a fixed length of 200bp, and aligned on the Human reference genome (hg19) with the splice mapper Tophat2. The different isoforms were then reconstructed from the aligned data and quantified using two different strategies (Cufflinks and R BioConductor package flipflop). All expected splicing events evidenced by Sanger were detected and the relative proportion of these anomalies was quantified. A number of alternative transcripts previously undetected by Sanger were also identified at a low frequency. This first result demonstrates the capabilities of RNA sequencing technics to assess the impact of VUS on splicing. The design of the PCR amplicons is crucial in order to extend this approach to more complex genes. This cost effective and simple approach can be easily implemented and has the potential to replace existing Sanger studies in a diagnostic context as the complex mixture of unstable, truncated, in frame and wild-type transcripts generated by splice mutations can be deciphered and quantified, hence allowing robust VUS interpretation and in turn reliable genetic counseling.

You may contact the first author (during and after the meeting) at