Resolving the genomic context of AMR genes from environmental metagenomes
We are developing novel bioinformatics and statistics tools to resolve the genomic context of AMR genes from short read shotgun metagenomics data. The direct sequencing of environmental DNA from the whole microbial community, metagenomics, has enabled the detection of AMR genes in environmental microorganisms (Li et al. 2015). Revealing key information on the location and prevalence of different genes (Amos et al. 2015) without the requirement to culture. This is a major step forward as the large majority of organisms cannot easily be grown in the laboratory.
The project is inherently inter-disciplinary, we will utilise informatics and statistics to extract biologically relevant information from an environmental data set. The pre- processing and assembly of the sequence data is computationally intensive and will only be possible through the use of cutting edge methods from computer science for storing and retrieving very large data sets. To resolve the ambiguity in the seeded assembly we will use methods from statistics, starting from a probabilistic model, which will be fitted with Bayesian inference methods, initially Gibbs sampling followed by more efficient approximations, principally stochastic variational inference. The interpretation of the results, the AMR genes and their genomic contexts, will depend on the expertise of Prof. Liz Wellington who is primarily an environmental microbiologist.
The project will provide critical information on which environmental organisms harbour AMR genes and exactly where in their genomes these genes are found. It is has been established that transfer of AMR genes between environmental organisms and pathogens occurs. It is highly likely that the environment provides a reservoir of genetic diversity that can maintain resistance in human associated pathogens. The first step in understanding the processes that generate the diversity of AMR genes in the environment and their transfer to pathogens is to obtain information on their abundance, their distribution across samples and crucially their links to the environmental organisms that harbour them. For instance, there are different types of HTEs, e.g. phages or plasmids, but we do not yet know their relative importance for the transmission of AMR genes.