Abstract 1418: Identification of recurrent regulatory mutations in breast cancer

• Published in: Cancer research (Chicago, Ill.) Ill.), 2018-07, Vol.78 (13_Supplement), p.1418-1418
• Main Authors: Cotto, Kelsy C., Danos, Arpad, Lesurf, Robert, Park, Morag, Griffith, Malachi, Griffith, Obi L.
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2018-1418

Since the identification of recurrent TERT promoter mutations in melanoma resulting in increased TERT expression, there has been increased interest in identifying recurrent regulatory non-coding mutations (Horn et al. 2013, Huang et al. 2013). Several studies have attempted pan-cancer analyses in order to identify these types of mutations, but often the results suffer from low coverage of regulatory regions or do not extend to breast cancer. While some breast cancer specific studies have identified some significantly mutated promoters and lncRNAs, they have often failed to incorporate transcriptome data to assess the impact and relevance of mutations on the expression of genes within tumors (Nik-Zainal et al. 2016). In order to address this, we assembled and generated a data set consisting of 458 breast cancer cases with matched tumor/normal pairs. This cohort consists of 22.4% luminal A, 19% luminal B, 16.4% HER2-enriched, 21% basal-like, 0.8% normal-like, and 20.4% unknown with regards to molecular subtype. This is important due to different breast cancer subtypes having dissimilar phenotypes and varying rates of gene coding mutations. This data set has a mix of whole genome, exome, transcriptome, and custom capture sequencing. We designed a custom capture reagent that covers regions assembled from regulatory databases, 5' untranslated regions, 500 bases upstream and downstream of transcription start sites, and 50,000 bases upstream and downstream of 178 genes that have been implicated as being important in breast cancer (Lesurf et al. 2016). While this custom capture region is similar in size to an exome, it has advantages over whole genome and exome sequencing, particularly with respect to coverage in GC-rich promoter regions. With these data, we predict that we will be able to identify novel, regulatory coding and non-coding drivers of breast cancer that would not be discovered without integrated analysis of the DNA- and RNA-seq data for each tumor. Instrument data were processed using the McDonnell Genome Institute somatic variant calling pipeline that includes 5 SNV callers and 3 indel callers. We then used these steps to filter variants: min. 20x coverage in both the tumor and normal sample, min. 2.5% tumor variant allele frequency, min. 3 variant supporting reads in the tumor sample, max. 10% variant allele frequency in the normal sample. We also filtered against gnomAD and a panel of normals. Rheinbay et al. 2017 identified recurrently mutated pr