Assessing the impact of non-coding mutations on breast cancer risk predisposition genes

  • Rayhaan Ali

Student thesis: Doctoral ThesisDoctor of Philosophy


Research efforts into women diagnosed with a Hereditary Breast or Ovarian Cancer (HBOC) has led to an improved understanding of the disease in some aspects, such as women with BRCA mutations. However, this is not the case for BRCAx patients (women diagnosed with a HBOC of unknown genetic cause), whose genetic predispositions are poorly understood. Indeed, the cause of between 50-60% of HBCs currently remains unknown, demonstrating the missing heritability of the disease. Furthermore, the role of mutations/variants in regulatory gene regions are not well understood. We have previously identified mutations/variants in regulatory regions of novel and known HBOC predisposition genes (many of which function within the DNA damage response (DDR) pathway), in a cohort of Northern Irish (NI) BRCAx patients. This study aims to define the NI BRCAx population from a clinico-pathological perspective and examine the impact of identified non-coding variants on risk predisposition gene expression, splicing and/or DNA repair pathway deficiency; and also model a number of these promoter variants for functional testing.

Initially, anonymised clinico-pathological data was gathered for the NI BRCAx cohort and linked to germline sequencing results to assess linked mutation/tumour phenotypes. To assess the impact of the non-coding variants on gene expression, targeted-panel RNA-Sequencing analysis in FFPE tumour samples from 96 patients with known gene regulatory region variants was performed. A publicly available RNA-expression gene signature, known as the DDR Deficiency (DDRD) signature, was utilised to assess for DNA repair deficiencies in these tumours. Identified promoter variants were also modelled in vitro using transcription reporter assays.

A total of 502 tumours were available for analysis. A large NI BRCAx clinico-pathological database has been developed that, to the best of our knowledge, is the first of its kind. The germline mutations these patients harboured and the corresponding tumour characteristics largely corresponded with other published BRCAx cohorts, illustrating the appropriateness and relevance of our cohort as a source for genetic studies assessing HBC biology/risk.

RNA-Seq analysis demonstrated that the majority of the non-coding variants in our cohort did not impact gene expression. However, 4 such variants in BRCA1, FANCG, RAD51D and MRE11A;ANKRD49 were shown to potentially have an effect on gene expression. These promoter variants are candidates for further investigating their role in breast cancer.

Methods for calculating DDRD scores from microarray data were adapted and, utilising RNA-Seq data from 1,211 breast tumours from TCGA, robust methods were designed for DDRD scoring RNA-Seq data inputs. These scores correlated strongly with those generated by independent groups. Upon DDRD scoring the selected sequenced group of NI BRCAx patient tumours, our cohort did not have a significant deficiency in DDR (only 9.4% were DDRD-positive) that could be identified using the assay.

Finally, transcription reporter assays suggested that specific substitutions in the RAD51D and YBX1 promoters appeared to affect downstream gene expression.

In summary, this study has shown that our NI BRCAx database is a unique and comprehensive resource. A small number of germline promoter variants may impact gene expression as observed in patient tumour samples, and provide variants for further study to help explain gene expression patterns of known breast cancer genes. This study also highlights that in the sequenced tumours, there was no apparent link between promoter variants and DDRD. Finally, data is provided to suggest a functional impact of two promoter variants that fall within sites with strong TF binding. Data generated from this study is also available for investigating splicing in the future. Together, findings from this study will help develop our understanding of non-coding variant contributions to breast cancer predisposition gene expression and risk.
Date of AwardDec 2020
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SponsorsNorthern Ireland Department for the Economy
SupervisorKienan Savage (Supervisor) & Stuart McIntosh (Supervisor)

Cite this