Results and Conclusions Development of a High Throughput Sequencing and Analysis Pipeline for West Nile virus Emmi Mueller1, Rebecca A. Halpin1, Nadia Fedorova1, Timothy B. Stockwell2, Danny Katzel2, Laura Kramer3, Suman R. Das1  1Virology Group and 2Informatics Department, J. Craig Venter Institute, Rockville, MD; and 3 Wadsworth Center, New York State Department of Health, Slingerlands, NY Abstract West Nile virus (WNV) is a member of the Flaviviridae family of enveloped, single-stranded RNA viruses with a linear, non-segmented, positive-sense ~11,000bp genome. WNV is an enzootic virus of birds and mosquitos, with humans and other mammals serving as spillover hosts, typically via mosquito bites. In humans, WNV is usually asymptomatic but ~20% of infected people develop a mild flu-like fever (West Nile fever) and ~1% can develop severe symptoms such as encephalitis or meningitis, which can be fatal. WNV was first isolated from a woman in Uganda in 1937 and first reported in North America in New York City in 1999. Since then, WNV has caused over 39,000 reported human disease cases in the USA contributing to over 1600 deaths. WNV continues to remain a formidable clinical and public health concern in the USA and worldwide. So far, no human vaccine or specific antiviral treatments for WNV infection are currently available. As part of the Genomic Center for Infectious Diseases (GCID), the J. Craig Venter Institute is developing a high throughput Next Generation sequencing and analysis pipeline to process ~650 complete WNV genomes from mosquitos, birds, and horses. The development of this pipeline will allow rapid and efficient sequencing and analysis of WNV genomes in order to compare and contrast the genetic diversity and evolutionary dynamics of WNV circulating within animal hosts. Furthermore, the sequencing data will allow us to assess temporal and spatial variation of WNV, as well as to identify selective pressures and genetic correlates of WNV disease transmission. Methods Reverse transcription (RT) and PCR WNV RT and PCR primers were designed manually off of a consensus of published WNV genomes to generate ~3kb overlapping amplicons across the genome. Extracted viral RNA was reverse transcribed with Super Script III reverse transcriptase (Thermo Fisher). For each of the eight amplicons PCR was performed using Phusion High-Fidelity DNA Polymerase (New England Biolabs). Bands were visualized and QCed on both 1% agarose gels and the QIAxcel Advanced System (Qiagen). 2. Sequencing and Assembly We successfully generated amplicons 1, 3, 5, and 7 (one amplicon per genome region) for 33 of the 93 samples. These amplicons were pooled based on concentration, Ion Torrent barcoded libraries were constructed, and the libraries were sequenced on the Ion Torrent PGM DNA sequencer (Thermo Fisher). Genomes were assembled using CLC Bio algorithms (Qiagen) resulting in 33 complete WNV genomes. 3. Analysis Phylogenetic analysis of the 33 genomes was performed using CLC Bio Genomics Workbench (Qiagen). A neighbor joining phylogenetic tree was generated using a K80 substitution model. Percent nucleotide identity table was calculated using a global alignment algorithm with a linear gap penalty of -2 and the Nucleotide 4.4 scoring matrix. Figure 5: Percent nucleotide identity (nt id) table comparing the 33 complete WNV genomes sequenced at JCVI, organized by year of collection. Lighter cells show a lower percent nucleotide identity. Results and Conclusions WNV amplification success varied, possibly due to RNA quality or PCR conditions. QIAxcel provided more precise amplicon QC results than gels with the benefit of concurrently quantitating the amplicons, thus saving time and resources. QIAxcel will be used exclusively in the future. Thirty three complete WNV genomes were assembled after a single round of sequencing, no gap closure was required. The 33 genomes are very similar in nucleotide sequence (98-99.9% nt id). WNV underwent gradual genetic drift between 2000-2002. In 2002-2003, two clades of WNV emerged. Both of these clades were maintained through 2014 (Figure 4). There is a general decrease in percent nucleotide identity between WNV genomes throughout time (Figure 5). 60528 60519 60520 60521 60522 60523 60524 60525 60526 60527 60528 60529 60530 5kb- -5kb 2.5kb- -2.5kb Figures 3a and 3b: Amplicon 3 for 93 WNV samples run on a 1% gel (Figure 3a) and 12 of the same samples (blue box in Figure 3a) run on the QIAxel (Figure 3b). Red dots on Figure 3a show amplicons that were pooled and sequenced. QIAxel was used to image and quantify the PCR products in a single step. QIAxcel is more sensitive to low concentration products, as shown for sample 60528 (green arrows). QIAxcel also requires 1/3rd of the volume and has a faster turn-around time as compared to a gel. Introduction The primary transmission cycle of WNV is between mosquitos and birds (Figure 1), with horses and humans serving as incidental hosts. A high throughput pipeline for sequencing WNV genomes from diverse hosts has not previously been developed at JCVI. We utilized 93 WNV-positive samples collected from mosquito pools in New York State between 2000-2014 to initially develop this pipeline. WNV PCR primers were designed with the goal of generating 8 overlapping amplicons across each genome (Figure 2). Two amplicons were designed in each of 4 genome regions so that there would be two amplicons covering each nucleotide and we would have a backup amplicon for each region in case one amplicon failed PCR or sequencing. Figure 3a Figure 3b Future Directions We have developed a preliminary pipeline for high throughput sequencing of complete WNV genomes. We will continue to optimize primers and PCR reaction conditions in order to improve the pipeline with a goal of sequencing and closing more than 93 genomes per sequencing run. We will sequence samples collected from various years and hosts (mosquitos, birds, horses) in New York State. Data will be reanalyzed to obtain a better picture of how WNV is evolving in multiple hosts over time. In total about 650 WNV samples will be sequenced and analyzed as part of the JCVI Genomic Center for Infectious Diseases. Figure 1: The transmission cycle of West Nile Virus. References & Funding Figure 1: http://www.mayoclinic.org/diseases-conditions/west-nile-virus/multimedia/west-nile-virus-transmission-cycle/img-20006044 Lanciotti,RS, et al. Science 17 December 1999: 286 (5448), 2333-2337. This project is funded by NIAID contract # U19AI110819 C -64 -432 -933 -2436 -3492 Amplicon 1 Amplicon 2 -4185 -4578 -6435 -6813 -7647 -10362 Amplicon 3 Amplicon 4 Amplicon 5 Amplicon 6 prM E NS1 NS2 NS3 NS4 NS5 Amplicon 7 Amplicon 8 5’ UTR 3’ UTR Figure 4: A neighbor joining phylogenetic tree demonstrating the phylogenetic differences between the 33 complete genomes sequenced at JCVI and the NY99 strain characterized in Lanciotti et al2. NY99 (shown in red as the outgroup) was isolated from a Chilean flamingo at the Bronx Zoo during the 1999 New York outbreak of WNV. Figure 2: The genome of West Nile Virus showing locations of genes and overlapping amplicons in four genome regions.