Understanding the genetic basis of disease endotypes in idiopathic pulmonary fibrosis

Idiopathic pulmonary fibrosis (IPF) is a rare, incurable disease of unknown cause characterised by progressive scarring of the lungs. The prognosis of IPF is poor with a median survival time of approximately 4 years and current treatment options are limited. The aim of the analyses in this thesis was to utilise genomic and transcriptomic data to improve the understanding of the pathogenesis of IPF, which could aid drug development and lead to improvements in treatments.

This thesis describes the first genetic analyses of the age at which IPF is first developed. First, genome-wide association studies were performed to identify common genetic variants that are associated with the age-of-onset of IPF. Following this, gene-based collapsing analyses were performed to investigate the role of rare genetic variation in the age-of-onset of IPF. These analyses highlighted some suggestively significant genes of potential interest as well as some important factors to consider when studying this phenotype.

A series of transcriptomic analyses were conducted to identify groups of IPF patients that could represent endotypes of the disease. New bioinformatics methods were utilised in these analyses to combine and cluster multiple datasets. This approach allowed for the largest transcriptomic cluster analysis in IPF to-date to be performed, which revealed three distinct groups of patients with IPF. These findings were consistent with the theory of multiple endotypes of IPF; significant differences in lung function and survival were found between clusters and gene enrichment analysis implicated metabolic changes, apoptosis, cell cycle and the immune system in the development of these potential IPF endotypes. Supervised machine learning was used to develop a gene expression-based classifier with the ability to assign patients with IPF to one of the three clusters. With further development, this classifier could be a useful clinical tool for outcome prediction and patient stratification in IPF.