FINGERPRINT: Fully Integrated Culturomic Platform for Rapid, High-throughput Microbial Identification and Characterisation

Lead Research Organisation: Queen's University of Belfast
Department Name: Sch of Biological Sciences

Abstract

Traditional microbiology is focused on growing cultures in the lab, but in recent decades this has changed. Developments in DNA sequencing technologies mean that it is possible to explore the genetic potential of a microbe without having to grow it, which has allowed the rapid study of entire communities of microbes in the human digestive system, soil, and oceans. While this has benefits, the traditional culturing of microbial communities ("culturomics") has seen somewhat of a Renaissance in recent years, as it has become clear that DNA sequencing cannot replace all other techniques. For example, while DNA sequencing will tell you which genes a microbe has, it does not tell you if these genes are functional, or when they will be used. It also cannot easily tell you what a new gene does, and many genes remain unstudied which could be important for medical or industrial applications. Finally, DNA sequencing is technically complex, slow, and costly, which restricts the amount of research which can be done.

Growing microbes isn't without its own problems, however. Microbial communities can be made of hundreds or thousands of species, and separating and growing all of these in the lab requires enormous amounts of work. To keep these cultures alive or to prepare an experiment they need to be transferred to fresh growth medium, which can be tedious and time-consuming for so many isolates. This creates a serious bottleneck to research and ties up valuable staff time. However, the process is simple: a sterile object is used to pick up the microbe and move it to fresh growth medium. This is a non-skilled and highly repetitive task, and so is ideal for automation using robotic systems. Colony pickers are machines designed with this purpose in mind, and are able to process thousands of microbes per hour, selecting them via computer vision systems based on size, shape, or colour. This allows many different experiments of large and complex formats to be set up using minimal staff time. For example, to find a gene involved in certain functions it is common to culture a "library" of thousands of strains, each with a different mutation. These mutants are then observed for any change in their behaviour which might reveal that they are linked to the function in question. Being able to manage thousands of these mutants automatically with a colony picker maximises the chances of identifying a gene which drives the process, using minimal staff time.

While the colony picker will help to separate, culture, and study these microbes, eventually we will need to know their identity. DNA sequencing is a well-established process for achieving this, but it is expensive and requires a lot of manual work for every microbe to be identified. The specialist machines and expertise needed mean it usually cannot be done in-house, and so precious samples must be posted - often abroad - to companies which specialise in this work. As a result, it can take several days for this process to be completed, and the costs often limit researchers to identifying just a few dozen microbes in a study. A new method of identifying microbes has been developed using a MALDI biotyper, which detects the unique fingerprint of proteins associated with different species. This allows much more rapid and inexpensive identification of large numbers of microbes, and so would perfectly complement the large libraries of microbes which would be cultured by the colony picking system.

Together, these two systems will enable us to greatly expand our research into how microbes create greenhouse gases in cow and sheep guts, how microbes cause disease in the human body, and how microbes in the natural world are able to survive and control the world around them. These are questions which drive the important societal questions of our time, and this platform will help us to answer them.

Technical Summary

Despite the explosion of modern, multi-omic technologies enabling us to study microorganisms with higher resolution than ever before, a major limiting factor to our understanding is poor culture collections and a lack of effort placed on culture work. Microbial physiology, behaviour, and interactions cannot be reliably predicted by genomic or metabolic reconstructions, and as such, an ability to culture, identify, and characterise microbes is essential to modern microbiology, despite the laborious nature of culturing processes. An additional fundamental requirement is the accurate identification of cultured microorganisms. Traditionally, this has involved gene sequencing, which can lead to significant time and cost investments. We propose to address these dual bottlenecks by combining a QPix 460 automated colony picking system with a Bruker Biotyper Sirius MALDI-ToF mass spectrometer into a platform for high-throughput, hands-off microbial handling and identification. The QPix 460 selects microbial colonies based on colour, fluorescence, size, shape, or halo production, and can inoculate them onto agar, or into microplates, and can process 3,000 colonies per hour without intervention. Cultures will be passed to the Biotyper for instant identification and characterization. This instrument can analyse ~200 isolates per hour; its current database consists of 3,239 species, and can create in-house databases of acquired spectra for further identifications. This enables the rapid and comprehensive identification of strains isolated and propagated by the QPix 460 system, at approx. 10% of the cost of sequencing. We will utilise this enhanced capability across eight work programmes led by QUB academic staff, covering novel insights into archaea, the human milk microbiome, microbial phosphorus cycling, the human airway microbiome, wastewater monitoring, links between microbiota and chronic respiratory disease, culturable rumen microbiota, and discovery of novel biocatalysts.

Publications

10 25 50