Identification of functional genetic variants affecting allele-specific chromatin conformation using Genome Architecture Mapping

You are here:

Applicant

Dr. Roland Schwarz

Max-Delbrück-Centrum für Molekulare Medizin (MDC)
Evolutionary and Cancer Genomics (Schwarz Lab)

https://www.mdc-berlin.de/de/schwarz

Summary

The identification of the effects of somatic and germline genetic variation is one of the most pressing challenges in genomics with far-reaching implications for health and disease. Traditional approaches have focused on the genome-wide association of variants with changes in gene expression (GWAS), which have revolutionised our understanding of human disease. Despite their success, GWAS approaches suffer from two major drawbacks. Firstly, while the effects of some genetic variants might directly be visible in expression levels, in many cases functionality of these variants is determined by intermediate phenotypes, such as epigenetic modifications and chromatin interactions, which are frequently ignored. Secondly, GWAS approaches are only feasible for common variants with a relatively high recurrence of more than 1% minor allele frequency in the population. In cancer, which is characterised by ubiquitous somatic variation with low to no recurrence, these approaches must rely on aggregation of somatic mutations in genomic windows, which severely limits detection power.  

The aim of this proposal is to develop new computational approaches for identifying functional genetic variation that affects chromatin conformation and promoter – enhancer interactions. To overcome sample noise and increase detection power we propose to derive allele-specific chromatin contact maps from Genome Architecture Mapping (GAM), a novel ligationfree genome-wide experimental technique developed by our collaboration partner Ana Pombo at the MDC. We will use GAM for the reconstruction of parental haplotypes to derive haplotype-specific contact maps and develop machine learning algorithms that predict differential chromatin contacts from their sequence context.  

This novel predictor will in turn be used to predict the effects of somatic genetic variants from the Pan-Cancer Analysis of Whole Genomes (PCAWG) consortium, in which we are currently investigating genometranscriptome interactions. Using these predicted chromatin contact changes as an intermediate phenotype, we will derive a novel way of associating multiple somatic variants in their sequence context with changes in allele-specific expression, thereby overcoming the limitations of traditional GWAS approaches in somatic association studies. This work will shed new light on the effects of somatic genetic variation on chromatin state, allow us to identify functional somatic regulatory variants and predict new regulatory drivers of cancer onset and progression.