Massively parallel sequencing for the microbiologist

Molecular biology techniques have revolutionised the diagnostic microbiology laboratory. In particular, the past decade has seen standardised nucleic acid sequencing methods applied to routine identification and typing of microorganisms. The full extension of this approach will see massively parallel sequencing (MPS; also referred to as next generation or high throughput sequencing) of samples, bringing with it new capabilities (e.g. complete community profiling in complex samples) and challenges. MPS is a different diagnostic paradigm requiring no prior hypotheses of the specific microorganisms in a given sample. While standard sequencing can detect non-culturable organisms in some circumstances (i.e. where a specific test is performed), MPS might enable hypothesis-free detection of all non-culturable microorganisms in a single assay. In addition, MPS has applications where the clinical picture is complicated or where standard diagnostic approaches have failed. MPS might also shed light on complex disease processes, particularly where disease involves the interaction of the host with a population of microorganisms. In the longer term, as sequencing technology and hardware for processing data become cheaper, the potential exists for rapid point-of-care testing of any microorganism.

The distinguishing feature of MPS is that it generates orders of magnitude more data in comparison with the older dideoxy chain termination method (‘Sanger’ sequencing). Depending on the platform (454, Illumina, SOLiD, Ion Torrent, etc.), human genome-scale amounts of data can be produced in a single experiment¹. There are common features across all platforms. First, the term ‘massively parallel’ refers to the large number of different reactions that are conducted on clonal DNA templates physically separated on microscopic beads, glass flowcells or micro-wells. In present generation technologies, the DNA to be sequenced is used as the template for the generation of a specific complementary product molecule. Similar to Sanger sequencing, the product is usually the result of ‘sequencing by synthesis’, a polymerisation reaction in which nucleotides are incorporated into a nascent strand using a suitable DNA polymerase. It is the real-time base-by-base read-out of the reaction products detected on large numbers of individual beads that generates the large volume of data. Platforms differ in the means by which the sequencing reaction product is detected, and this is an important factor in the variation in the speed and cost per base. The next step in this technology is the so-called ‘third generation’ single molecule sequencing (SMS) that literally uses a single nucleic acid input molecule². While there are few commercially available products or services offering SMS, the potential advantages in read length, cost and sensitivity mean that demand for this technology will be great in the coming years. At the forefront of SMS are Pacific Biosciences, who have a commercially available sequencer and Oxford Nanopore who, at the time of writing, have not yet reached the market.

The most common infectious disease scenario is when a patient presents to a clinician with a particular set of symptoms. To arrive at a diagnosis, the clinician draws on their knowledge and experience to narrow down the choice of potential tests from a large pool. The result of the test is usually either detection or isolation of the microorganism, or detection of an immune response indicating that the microorganism is, or has been, present. Central to this model is the selection of the correct test or the performance of a range of tests covering a number of possibilities, with the associated costs, regardless of the number of negative results. This places significant emphasis on the clinician’s experience and expertise. MPS offers the intriguing possibility of a single test capable of the hypothesis-free detection of any microorganisms present, and the potential to save time and money by avoiding unnecessary testing. In addition, for some diseases a more complicated understanding is emerging, particularly in the case of gut microflora, of the role of microbial communities and their interaction with the host in determining health and disease. MPS is at the forefront of this research and is the most applicable technology to generate a profile of a microbial community. A popular approach in the research world is to amplify and sequence ribosomal RNAs using conserved primer sets. Using so called ‘deep sequencing’, the technique is capable of not only identifying microorganisms, but also providing quantitative information as well. The application of deep sequencing has been central in revealing a number of changes in the gut microbiota associated with diseases such as allergies, obesity, celiac disease, inflammatory bowel diseases³ and susceptibility to viral infections⁴. This approach is being applied to oral, nasal, skin and genital microbial communities and is expected to illuminate many other diseases involving complex interactions between microbes and their human hosts⁵. This opens the possibility of new therapeutic approaches to encourage ‘healthy’ microbial communities.

MPS/deep-sequencing techniques are also applicable to viruses. At the research level, MPS can shed light on populations of related genomes such as quasi-species present during HIV infection that enable the virus to rapidly adapt to host immune and antiviral drug pressures. Additionally, during the outbreak of a new emerging pathogen or infections with rare and unusual agents, MPS can be used to rapidly obtain sequence information that can facilitate the development of diagnostic reagents^6–13. A good example of this is the recent human beta-coronavirus in late 2012, in which genome sequence information was made available using MPS in a matter of weeks¹⁴ and might actually be possible within days using the most recent machines. Many microorganisms in the environment are not able to be cultured. So another advantage of MPS over traditional methods is that genome information can potentially be obtained directly from the sample without prior culture.

In the not-so-distant future we can imagine the following scenario in a doctor’s surgery or hospital ward. A febrile traveller presents and a throat swab is taken. The sample is placed in a hand-held diagnostic device that processes the sample through its microfluidics, extracting nucleic acids. Single molecule sequencing occurs, rapidly generating gigabases of data that are analysed by an in-built microprocessor, which compares it to a database for matches with known microorganisms. A best-match reveals a high probability of an exotic viral infection. Combined with wireless and an Internet connection, there is the possibility of instant disease notification of the infection to relevant public health authorities. For less serious infections, the same technology could be used for real-time global disease monitoring. There are obviously many important quality control and ethical issues to be solved first. However, the main challenge to the realisation of such a scenario is, surprisingly, not primarily technical. Sadly, we might all be out of a job sooner than you think!