London Calling

In this talk I will cover the highs and lows of being part of the Oxford Nanopore MinION Access Programme. Our laboratory joined the MAP programme in May 2014. Soon afterwards we published the first nanopore read (viewed >4000 times on FigShare) and we released a toolkit for analysing MinION squiggle data (poretools). We piloted the use of nanopore in an large hospital outbreak of Salmonella and demonstrated the utility of real-time sequencing. In September we released a large reference dataset of E. coli K-12 which was used as a reference dataset for bioinformatics algorithm development, and we managed to produce a de novo assembly from these data alone. Now we are turning our attention to using the MinION nearer the patient and in more remote locations for the purposes of outbreak epidemiology and metagenomics diagnostics. Challenges include shrinking the molecular laboratory into a backpack, reducing DNA inputs and developing new software for rapid data analysis.

Here we present our experiences working with MinION device. We discuss the performance, tools and algorithms available for downstream analysis; including alignment methods and parameters. Our interest in the MinION device is its ability to generate long reads for complex genome assembly. Using our hybrid error correction approach called Nanocorr, we filtered the reads to optimize length and quality and corrected errors in the Oxford Nanopore reads using high identity MiSeq reads collected from the same strain. We generated 28X coverage of the genome with reads >7kb. The process improved the raw nanopore reads with a mean accuracy of ~65% to an average per base identity of 97%. The corrected reads were then assembled using the Celera Assembler which now supports reads up to 500kbp. The assembly had an average identity of 99.78% and an N50 value of 585kb. We also highlight the utility of the corrected nanopore assembly by showing that the long reads generated by the MinION are superior to MiSeq short reads in detecting complex genomic elements larger than 1300bp in length.

River waters worldwide are impacted by disease-causing agents including bacteria, protists, flatworms, viruses, and harmful algae that derive from domestic sewage and farm runoff, and/or are emergent due to nutrient pollution and climate change. We are determining the ability of MinION long read metagenomic data to identify pathogens that otherwise would require numerous individual assays to detect. We also are assessing the utility of long read data to circumvent false positive matches from shared conserved sequences in short read data (the customary method used to generate metagenomic data), and the extent to which the phylogenetic diversity inferred from the long reads mirrors the profiles obtained by the sequencing of rDNA amplicons. Our preliminary results suggest that MinION data can accurately predict the composition of synthetic metagenomes. Presently, we are assessing how the diversity detected by MinION from environmental metagenomes compares to individual PCR-based detection of pathogens and we are testing ways to use this platform to monitor the distribution of organisms over space and time. This metagenomic approach is a novel application of MinION that can improve public health associated with environmental exposure to biological agents and that has the potential for worldwide application, especially in developing nations where public health genomics is urgently needed.

In my talk, I will discuss my collaboration with Nick Loman's lab to develop de novo assembly methods for MinION data. We have built a pipeline to error correct nanopore reads using partial order graphs and the corrected reads are subsequently assembled using the Celera Assembler. The main focus of my talk will be on the post-assembly consensus calling stage, which uses the electrical current signal emitted by the MinION. I will describe in detail the probabilistic model we have developed for MinION signal data and our software tools for working in signal space. Our approach assembled the E. coli genome into a single contig with 99.4% accuracy after signal-level polishing.

Until today, the investigation of the biological diversity of many parts of the world has been obstructed by the inability to perform field based DNA analysis, especially in those regions where the biological richness is higher, like the countries in the inter-tropical area. Some years ago, Edward Osborne Wilson wrote that the possibility to sequence the DNA of organisms in the most remote places of the planet, making possible the identification of any organism, one day will allow a sensational step ahead in the study of biodiversity. On May 4th, 2015 a team of Italian researchers performed the first genetic tests on the slopes of an extinct volcano in the forests of southern Tanzania, making Wilson’s dream a reality. Using the MinION device, they sequenced the DNA of a wild frog species and transmitted the raw data via a standard smartphone to Oxford Nanopores cloud based application, Metrichor. Metrichor performed a custom automated workflow that allowed the team in Tanzania to determine in real time the identification of the species of frog that they had just sequenced. The data was simultaneously sent to their colleagues at the MUSE of Trento where they were able to participate in the discovery.

The diagnosis of infectious diseases by culture takes at least two days: one to grow the bacteria and then, at best, one to identify pathogens and test their antimicrobial susceptibility. Meanwhile the patient is treated empirically, which often results in inappropriate treatment. A paradigm shift in diagnostics technology is required, to allow the development of a universal diagnostic which can detect any pathogen, known or unknown. We are developing the first culture independent non-targeted NGS based methods for the routine diagnosis of clinical syndromes such as sepsis and urinary tract infections. NGS based infectious diseases diagnostics represents a disruptive advance in the field, providing rapid, comprehensive molecular diagnostics. The biggest challenges to successfully applying NGS in medical microbiology are (1) the vast quantity of host vs pathogen nucleic acid present (e.g. blood) and (2) turnaround time to results. We are combining novel pathogen DNA enrichment techniques with rapid MinION sequencing to make NGS based infectious diseases diagnostics a reality.

Haplotypes are often critical for the interpretation of genetic laboratory observations into medically actionable findings. Current massively parallel DNA sequencing technologies produce short sequence reads that are often unable to resolve haplotype information. Phasing short read data typically requires supplemental statistical phasing based on known haplotype structure in the population or parental genotypic data. Here we demonstrate that the MinION nanopore sequencer is capable of producing very long reads to resolve both variants and haplotypes of HLA-A, HLA-B and CYP2D6 genes important in determining patient drug response in sample NA12878 of CEPH/UTAH pedigree 1463, without the need for statistical phasing. Long read data from a single 24-hour nanopore sequencing run was used to reconstruct haplotypes, which were confirmed by HapMap data and statistically phased Complete Genomics and Sequenom genotypes. Our results demonstrate that nanopore sequencing is an emerging standalone technology with potential utility in a clinical environment to aid in medical decision-making

Nanopore sequencer, MinION, has enabled sequencing analysis without pre-installation of expensive conventional sequencers or pre-requisite of specific skills in biological experiments. Even electric supply is not always necessary, by connecting MinION to a laptop PC. These features of MinION have opened the opportunity to enable precise genotyping of pathogens in tropical diseases in a developing country even in its filed areas. In this study, we attempted genotyping Dengue viruses regarding their serotypes (types 1-4). We directly used serum samples of Indonesian Dengue patients, from which viral genomes were directly amplified by the reverse-transcription-LAMP method in an isothermal reaction condition. We directly used the amplified templates for MinION sequencing allocating one flow cell per sample. We found, although the overall sequencing quality was low (70% sequence identify to the reference genome and the quality value of QV= 5 on average), thereby obtained sequence data could discriminate different serotypes of the viruses, whose genome sequences were diverged with the sequence similarity of 70%, with the overall accuracy of 98%. To further examine whether MinION sequencing can be also applied for detecting SNVs, we conducted genotyping of presumed drug resistance-causing SNVs in malaria parasites, Plasmodium falciparum. We similarly subjected ten PCR amplicon-mixes covering these SNVs to the MinION sequencing. In spite that the sequence alignments generated by a Smith-Waterman-based program, SSEACH showed that the average sequence identity was 65%, we found that the mutations at a particular position could be called by the accuracy of 85%, when all the reads covering the corresponding positions were collectively evaluated. Taken together, we provide the first simple experimental and analytical MinION sequencing procedure, which can be easily followed in a developing country to effectively genotype pathogens of tropical diseases.

Determining the full-length genome sequences of viruses during disease outbreaks such as the ongoing Ebola virus outbreak in West Africa, which is of unprecedented scale with about 24,000 cases and 10,000 deaths as of March 2015, can provide important information about virus evolution, and ensure the continued efficacy of molecular diagnostics and sequence-based countermeasures such as siRNAs. However, transport of samples out of affected countries and be logistically and politically challenging, and establishing conventional Sanger or Next-Generation sequencing technologies, which are heavy and of larger foot-print, in outbreak areas poses significant logistical and technical hurdles. The MinION sequencing device represents an extremely attractive alternative due to its small size and robustness as compared to other platforms, and we developed sequencing protocols allowing determine the full-length sequences of Ebola viruses using this technology. We will share our experiencees developing these protocols, as well as our experiences field-testing the device under outbreak conditions.

Are 1-D and 2-D gel-electrophoresis the best we can do in routine protein analysis? This talk demonstrates that by moving proteins through nanopores, it is possible to separate these proteins transiently from other macromolecules in solution. This unique feature has the potential to enable instantaneous and unambiguous characterization of the one protein in the pore. We show that monitoring the rotational dynamics of this protein inside the pore then makes it possible to determine several excellent protein descriptors simultaneously, including the protein's volume, shape, charge, dipole moment, and rotational diffusion coefficient. The resulting 5-dimensional protein characterization may open the door to real-time protein characterization, identification, and quantification on a single molecule level in complex mixtures. We think that this approach has the potential to be more sensitive, specific, quantitative, reproducible, and rapid than gel-electrophoresis while providing additional protein descriptors and requiring only minute sample sizes. This technique may replace the ubiquitous use of gel-electrophoresis for protein analysis all around the world with a chip-based approach to instantaneous protein counting that may also allow single protein sorting.

Nanopore sequencing represents a paradigm shift in DNA sequencing, and today it is the only sequencing technology that measures an actual single molecule of DNA. The Oxford Nanopore MinION is the world’s first mobile DNA sequencing device, measuring 4” in length and powered by the USB port of a laptop. At the University of Edinburgh, we have been involved with the MinION access programme since the beginning and have used the platform to sequence a number of small- to medium- sized genomes, and have used the MinION data both for variant calling and scaffolding of genomes. I will describe the results of these analyses. We have also developed software to help organise and analyse MinION data, and we will describe the development path for this software, which includes enabling it for mobile, in-field pathogen detection. Finally, I will present a future vision for the MinION platform, and nanopore sequencing in general.