Personal tools
You are here: Home Events ANC Workshop: Giulio Caravagna, Chair: Lynda Webb

ANC Workshop: Giulio Caravagna, Chair: Lynda Webb

— filed under:

  • ANC Workshop Talk
When Jun 27, 2017
from 11:00 AM to 12:00 PM
Where IF 4.31/4.33
Add event to calendar vCal


Detecting repeated evolution in cancer from multi-region sequencing datasets using transfer learning.




Carcinogenesis is an evolutionary process driven by the accumulation of genomic aberrations in somatic cells. Recurrent sequences of genomic events in cancer, both between and within patients, reflect repeated evolutionary trajectories that are important for anticipating progression and treatment resistance. Multi-region sequencing and phylogenetic analysis of human cancers allow inferring the partial temporal order of genomic changes within a patient’s tumour. However, phylogenetic trees from different cancers appear very diverse and recurrent evolutionary trajectories are hard to identify as they are obscured by the underlying complexity and stochasticity of the evolutionary process, hence the need for rigorous and sensitive tools to detect repeated evolution in cancer. I will show our recent work where we propose to use Transfer Learning (TL), applied to multi-region sequencing datasets (n=160 patients) to detect repeated evolutionary trajectories in lung, breast and renal cancers. Our method overcomes the stochastic effects of evolution in each individual patient, highlighting hidden recurrences of evolutionary trajectories in patient cohorts. We identified novel evolutionary subgroups characterised by specific recurrent trajectories that reproduce in large single-sample cohorts (n=2,341) and have prognostic value. Our method provides a novel patient classification measure that is grounded in the cancer evolution paradigm, and allows inferring access to evolutionary trajectories during tumourigenesis. This study has implications for our ability to anticipate malignant evolution.