Final Report Summary - SMARTBAYES (Intelligent Stochastic Computation Methods for Complex Statistical Model Learning)
The ERC project SmartBayes created a significant number of novel statistical models and computational inference methods for solving important scientific and technological problems in several application areas such as computational biology, forensics and pathogen evolution and transmission. Our flagship methods for inference about intractable statistical models go significantly beyond the state-of-the-art and possess considerable potential to initiate novel research directions in this area and to solve future research problems by enhanced scalability and accuracy. The Bayesian population genomic methods developed in SmartBayes have recently become gold standard in large-scale studies of bacterial evolution, where no other Bayesian model-based methods are currently capable of reliably handling hundreds to thousands of whole-genome sequences. Used in collaboration with the leading scientists in pathogen research, these methods have enabled several significant advances in understanding pathogen transmission, the evolution and horizontal transfer of resistance and virulence elements, and their interplay with ecology. The impact of this research stems from the future potential to use the generated understanding to reduce burden of infectious diseases in the human population.