Computational and Disease Genomics

We are trying to find out the DNA letter changes that cause disease, more precisely complex disease and cancer. A DNA change that alters someone’s risk of complex disease is hard to pinpoint because it is present among a large set of other changes that do not alter risk: it is a needle in a haystack. We are taking an unconventional approach that presupposes that changes in how a specific protein binds DNA alters disease risk. At the beginning of the study we do not guess which disease this is, but instead allow the human DNA data to reveal this. Our cancer research compares changes to both the DNA and RNA (a read-out of DNA) of single cells, and builds up a picture of how cancer is triggered and how its cells evolve over time. We hope to study multiple cancers that co-occur, and so find out whether they compete or cooperate. Finally, we apply our expertise in evolution to predict the functions of proteins that are altered in genetic disease.

We seek to advance the molecular mechanistic understanding of genetic disease and transcriptional regulation. To do so we take advantage of our recent analytical and technological advances in understanding noncoding sequence functionality, sequencing the DNA and RNA of single cells, and identifying distantly related homologues. We have three main aims:
• Determine the causal contribution made by variable transcription factor binding to an individual’s susceptibility to complex disease
• Tracing the competing lineages of genomic/transcriptomic crisis from very early oncogenesis
• Determine the function and evolutionary history of new protein domains.

Our first hypothesis is that variation of trait susceptibility occurs via a causal effect on gene expression arising from genetically or environmentally altered transcription factor (TF) binding. We are using Mendelian randomization approaches that predict causal relationships and do so for genetic, rather than purely environmental, contributions to disease risk. We study a TF for which there is existing genetic (i.e. causal) evidence of contribution to complex disease and whose activity can be easily and safely modulated therapeutically. We seek to predict the benefit of this therapy for persons according to their polygenic risk for specific complex diseases that we will identify objectively. Our selected factor is the vitamin D receptor (VDR), a TF of the nuclear receptor family that functions as a heterodimer with retinoid X receptor (RXR), chosen because supplementation of vitamin D (cholecalciferol or ergocalciferol) at pharmacological doses is common and is safe. The approach that we are developing will, however, be applied later to other factors in additional cell types.

Secondly, we will identify competing lineages of genomic/transcriptomic crisis from very early oncogenesis. At the origin of all cancers is a chance event occurring in a single cell. This somatic mutational event tips the balance in favour of the cell’s progeny becoming oncogenic. We will first identify such tipping point driver mutations lying near to the root of pre-neoplastic cell lineages by sequencing both the DNA and RNA from single cells. We then will determine whether the subsequent genotypic evolution of this cellular lineage is reproducible. By sequencing both the DNA and the RNA from the same cells we will trace the cells’ phenotypic trajectory as they acquire successive mutations down their lineage. Finally, we will assess competition among co-existing clones.

Thirdly, we have predicted and are experimentally investigating novel domains found within the
30 subunits of human Mediator. The approaches that we have developed are also being applied to other large complexes of central importance to human disease.