Accurate in vivo population sequencing uncovers drivers of within-host genetic diversity in viruses

Abstract

1Mutations fuel evolution and facilitate adaptation to novel environments. However, 2 characterizing the spectrum of mutations in a population is obscured by high error rates of next 3 generation sequencing. Here, we present AccuNGS, a novel in vivo sequencing approach that 4 detects variants as rare as 1:10,000. Applying it to 46 clinical samples taken from early infections 5 of the human-infecting viruses HIV, RSV and CMV, revealed large differences in within-host 6 genetic diversity among virus populations. Haplotype reconstruction revealed that increased 7 diversity was mostly driven by multiple transmitted/founder viruses in HIV and CMV samples. 8Conversely, we detected an abundance of defective virus genomes (DVGs) in RSV samples, 9including hyper-edited genomes, nonsense mutations and single point deletions. Higher 10 proportions of DVGs correlated with increased viral loads, suggesting increased cellular co-11 infection rates, which enable DVG persistence. AccuNGS establishes a general platform that 12 allows detecting DVGs, and in general, rare variants that drive evolution. 13 2016; Wang, et al. 2017); error reduction by overlapping paired reads in paired-end sequencing 46 (Chen-Harris, et al. 2013;Schirmer, et al. 2015;Preston, et al. 2016); and usage of improved 47 polymerases (Imashimizu, et al. 2013). However, most experimental methods described above 48 are designed for samples with high biomass and are inapplicable for sequencing of clinical 49 samples, where the biomass may be extremely low. Furthermore, these experimental protocols 50 may introduce their own artifacts to the sequencing process (Lou, et al. 2013; Brodin, et al. 51 2015). On the computational side, it has been suggested that well-established variant callers do 52 not perform well on clinical virus samples (McCrone and Lauring 2016). Here, we sought to 53 develop a simple and rapid approach that can tackle the problem of accurate sequencing of 54 clinical samples, and applied it to study the early stages of virus infection. 55We describe AccuNGS, a simple yet powerful approach for accurate population sequencing and 56 bioinformatics variant calling. We extensively optimize all stages of the method to ensure high 57 accuracy and maximal yield. We use AccuNGS to perform in-depth sequencing of 46 samples 58 from three different major human pathogenic viruses: human immunodeficiency virus (HIV), 59 respiratory syncytial virus (RSV), and cytomegalovirus (CMV), all sampled during the acute 60 infection stage. We compare the within-host genetic diversity among and within different virus 61 populations, and find patterns characteristic of each virus. We demonstrate the role of multiple 62 transmitted/founder viruses as major contributors to the genetic diversity in HIV and CMV. 63 Furthermore, we identify and quantify the impact of various host editing enzymes on the 64 mutational spectrum of viral genomes/populations in vivo. Intriguingly, we find that RSV 65 samples bear much higher levels of potentially defective virus genome (DVGs) ...

Full genome viral sequences inform patterns of SARS-CoV-2 spread into and within Israel

“…Sequencing reads were trimmed using pTrimmer, a multiplexing primer trimming tool (Zhang, et al 2019), and then aligned to the reference genome of the SARS-CoV-2 (GenBank ID MN908947) using our AccuNGS pipeline (Gelbart, et al 2019), which is based on BLAST (Altschul, et al 1997), using an e-value of 10 -9 . The pipeline allows for consensus determination and variant calling.…”

Direct sequencing of RNA with MinION Nanopore: detecting mutations based on associations

“…Both populations were derived by fifteen serial passages performed at 37 • C (denoted as A and B) as part of an evolve-and-resequence experiment (Methods). We first performed deep sequencing of both populations at passage 15 with the Illumina MiSeq platform (34). This revealed several segregating mutations (Figure 1), some of which shared similar frequencies.…”

“…Library preparation was performed according to our lab's AccuNGS sequencing protocol (34) with some modifications: The reverse transcription reaction was performed using SuperScript ® III Reverse Transcriptase (Thermo Scientific), the reaction was performed on 500 ng of RNA phage with 1 l of dNTP mix (10 mM), 1 l of 3R primer (TG GGTGGTAACTAGCCAAGCAG) (10 M) and 10 l of sterile distilled water. The mixture was incubated for 10 min at 65 • C followed by incubation on ice for 7 min.…”

“…A total of 417,000, 105,000 and 400,000 reads were produced for the MinION-p15A, MinION-p15B and the control runs, respectively. In order to map the reads to the MS2 reference genome we ran our computational pipeline (Materials and Methods) (34), which infers the proportion of each point mutation (A, C, G, T or '-') at each position in the genome. Over 97% of the reads were mapped to the reference, yet often sequencing terminated before it reached the 5 end in both the evolved and control populations (Supplementary Figure S1).…”