Mapping Homologous Recombination Genes through Omics Data Integration

Introduction

Mutations in homologous recombination repair (HRR) genes trigger numerous genetic disorders. In hereditary breast and ovarian cancer (HBOC), these mutations are the drivers underlying ~30% of cases. The genetic cause of the remaining 70% is unknown, but molecular evidence suggests the involvement of uncharacterized HRR genes.

Material and method

To generate a complete cohort of genes that are functionally related to the HRR, we integrate 24 different genetic screens and databases into a database (HRRbase) that maps the interaction of every gene with the HRR pathway. We use machine learning to rank the genes and create a list of HRR genes.

Results and discussion

Naïve Bayesian classifier analysis of this database yields 435 genes predicted to be highly involved in HRR. We molecularly validate a subset of these predictions. We found mutations in two genes from our list in breast cancer patients.

Conclusion

The genes coming out of our analysis may play a significant role in HRR related genetic disorders and HBOC, potentially yielding new targets for therapeutics.