Analysis of 'next-generation' sequencing data
We are interested in all aspects and applications of next generation sequencing technologies from quality control to data visualization. We have developed methods for identifying copy number changes in DNA-seq, identifying binding sites in ChIP-seq, detecting differential expression in RNA-seq, and identifying small RNAs. We generate our own data as well as collaborating with other research groups, and have particular interests in combining multiple data types.
Evolutionary approaches to cancer
Since the 1960s, population growth models have been developed to explain the growth dynamics of tumours, but have often failed to capture the complex spatial and temporal mechanisms involved in the disease. We develop novel spatial stochastic models to study, for example, the cancer stem cell hypothesis, and support these by generating large amounts of data on cancer heterogeneity using high-throughput sequencing.
Statistical methods for microarray data
Microarray technologies continue to be an important tool in cancer studies. We continue to provide open source computational tools for microarray analysis and have developed over 20 software packages which receive tens of thousands of downloads a year. A long-running theme has been the development of methods relating to Illumina’s BeadArray technology, for which we have also led a concerted reannotation effort in order to reduce misinterpretation of such studies
Design of experiments
We conduct studies to identify the best platforms to use for particular questions, and to cope with the varying quantities and qualities of material that it is typical to encounter in clinical cancer studies. We consider aspects of sample-size, sequencing depth and time-point selection in next-generation sequencing studies, and look for practical design solutions to balance theoretical optimality with robustness to errors and sample failure.
Statistical genetics and epigenetics
Naturally we are interested in the association of genetic or epigenetic markers with risk of disease, nature of diseases, and rate of disease progression. A specific example is the development of methods for the analysis of association studies that allow for predictors to be acting in combination. We are interested in the application of approaches from population genetics to the study of tumours.
This project is led by Rebecca Fitzgerald, which seeks to sequence 500 Oesophageal Adenocarcinoma tumour/normal pairs, and in which the group is leading the analysis effort.
We are also collaborating with David Neal, Ros Eeles and Colin Cooper on aspects of the UK Prostate ICGC initiative.
We have been collaborating with Dr Colin Watts, Dr Christina Curtis and Dr John Marioni in a study of intra‑tumour heterogeneity in glioblastoma.
We are part of the RADIANT consortium (“Rapid development and distribution of statistical tools for high-throughput sequencing data”) coordinated by Magnus Rattray. This is a European project that aims to bring the most advanced statistical tools for high-throughput sequencing into the mainstream of biomedical research
Within the Institute, we have notable collaborations with the Narita, Winton, and Caldas laboratories.