Microbial Outbreak Surveillance System: Establishing sequencing efforts with automated analytical pipelines for outbreak surveillance
In the Genomic Epidemiology Research Group at The Technical University of Denmark, we are researching and developing methods for establishing sequencing efforts in remote, decentralized locations all over the world. By delivering an all-in-one solution, which is comprised of in-house developed back-end workflows, a laptop with a powerful GPU, an Oxford Nanopore MinION sequencer and a user-friendly graphical user interface, we are enabling genomic sequencing in locations where it would otherwise not be possible.
Currently, we are focusing on enabling the sequencing and analysis of bacterial isolates from clinical settings. The vision is that one day every local health clinic or hospital could have a sequencing set up in which a non-bioinformatician/microbiologist could perform the library preparation, sequencing, and subsequent analysis after only a week of training. We are striving to achieve this vision by automating as much of the process as possible.
Post-sequencing, it only requires the user to fill in relevant meta data about the isolate, and then with the press of a button, the input data are analyzed, and the results compiled into an on-site usable PDF report with clinically relevant information. Additionally, important information such as virulence genes, AMR genes and phylogeny is automatically shared with a centralized cloud server. This allows for synchronization between each hospital/clinic, thus enabling real-time epidemiological decision to be taken, which might prevent future outbreaks of dangerous pathogens.
Naturally, this does raise the question; Why do we not just sequence and send all of the data to the cloud? Correctly assumed, this would allow for much, much simpler workflow setups for the analytical pipeline. This might be a viable solution in highly developed countries where high-speed internet is widely available, and eventually this is likely the scalable solution that should be implemented for a real-time global solution. However, as we have experienced during our research expeditions to Africa, in large parts of the world, sufficient internet capacity to send the required quantitates of data produced by the Oxford Nanopore sequencers is simply not available. One thing is to send the base-called reads – which would require an onsite GPU in the laptop anyways – but to possibly send the fast5 data (which is much, much larger) would be nearly impossible. Therefore, we instead focus on local sequencing, local base-calling, local analysis and then sharing of the compiled analytical results.
In the future, we also aim to implement similar workflows for analysis of viruses and metagenomic samples.
Malte Hallgren’s presentation