EXPLORING ALIGNMENT-FREE SEQUENCE COMPARISON METHODS TO ELUCIDATE PATTERNS OF EVOLUTION AND HETEROGENEITY IN LONGITUDINAL GLIOMA PATIENT COHORTS

Longitudinal glioma patient samples can reveal critical information relating to gliomagenesis, tumour evolution and therapy resistance. Previous studies examining tumour heterogeneity and evolution in these cohorts have focused solely on the limited information retained after applying stringent alignment and variant calling pipelines. Here, we propose the use of alignment-free methods to obtain an unsupervised view of the relationships between longitudinal tumour samples. Alignment-free (AF) sequence comparison is a novel methodology that has been heavily implemented in protein sequencing comparison and assessing evolutionary relationships between organisms but has not yet been applied to cancer research. Unsupervised AF methods can result in shorter computational times and may offer wider information for studying tumour heterogeneity and evolution in cancer research. In this study, we have identified a cohort of longitudinal glioma samples (Johnson et al.) for which we have whole exome-sequencing for germline, multiple spatial samples from an initial grade II glioma and spatial samples from a subsequent recurrent tumour for each patient. We developed an AF software to produce unrooted neighbour-joining trees for each patient, for which a matched least-square minimum-evolution tree produced from non-synonymous, somatic SNPs present in each of the samples is available from the original publication. Results demonstrate clear similarities when comparing the results from AF analysis to the trees produced using alignment-based approaches, however, clear and possibly fundamental differences were identified. These may be a result of the compounded effect of unassessed structural variants, larger indels or passenger synonymous variants which are also contributing to the tumours mutational landscape and therefore effecting the evolutionary pathway of gliomas. We hypothesise that this presents an opportunity to identify early and late events in evolution as well as highlight potential therapy-resistant subclones of the initial tumour which then give rise to a subsequent recurrent tumour.