Using high-throughput sequencing to leverage surveillance of genetic diversity and oseltamivir resistance: a pilot study during the 2009 influenza A(H1N1) pandemic

Abstract

Background: Influenza viruses display a high mutation rate and complex evolutionary patterns. Next-generation sequencing (NGS) has been widely used for qualitative and semi-quantitative assessment of genetic diversity in complex biological samples. The "deep sequencing" approach, enabled by the enormous throughput of current NGS platforms, allows the identification of rare genetic viral variants in targeted genetic regions, but is usually limited to a small number of samples.

Methodology and principal findings: We designed a proof-of-principle study to test whether redistributing sequencing throughput from a high depth-small sample number towards a low depth-large sample number approach is feasible and contributes to influenza epidemiological surveillance. Using 454-Roche sequencing, we sequenced at a rather low depth, a 307 bp amplicon of the neuraminidase gene of the Influenza A(H1N1) pandemic (A(H1N1)pdm) virus from cDNA amplicons pooled in 48 barcoded libraries obtained from nasal swab samples of infected patients (n = 299) taken from May to November, 2009 pandemic period in Mexico. This approach revealed that during the transition from the first (May-July) to second wave (September-November) of the pandemic, the initial genetic variants were replaced by the N248D mutation in the NA gene, and enabled the establishment of temporal and geographic associations with genetic diversity and the identification of mutations associated with oseltamivir resistance.

Conclusions: NGS sequencing of a short amplicon from the NA gene at low sequencing depth allowed genetic screening of a large number of samples, providing insights to viral genetic diversity dynamics and the identification of genetic variants associated with oseltamivir resistance. Further research is needed to explain the observed replacement of the genetic variants seen during the second wave. As sequencing throughput rises and library multiplexing and automation improves, we foresee that the approach presented here can be scaled up for global genetic surveillance of influenza and other infectious diseases.

Figure 1. Distribution of A(H1N1)pdm influenza cases during May-Nov, 2009 period.

A) Overview of the State of Morelos, Mexico. Bars represent the number of cases according to municipality during the first wave (weeks 17–32; blue section of each bar) and during the second wave (weeks 33–47; orange section of each bar). Each individual case is represented in green (haplotypes not associated with oseltamivir resistance) and in red (either the H275Y or the S247N mutations). B) Zoom in of Cuautla city where three clustered cases of the S247N mutation were found.

Figure 2. Sequence coverage plots of the sequenced NA amplicon according to barcode (A) and according to picotiterplate region (B).

Positions of the NA gene are shown in X axis, whereas the number of reads is shown in Y axis. Less variability in read number was obtained according to MID than according to sequencing region, suggesting that the source of variability is during the emPCR amplification.

Figure 3. Genetic diversity and frequency of genetic variants identified in the 717–982 region of the NA gene of Influenza A(H1N1)pdm virus.

A) Phylogenetic tree based on the average distance of the percent identity between translated sequences from the 30 identified haplotypes. The group I and II represent unique haplotypes with non-synonymous changes. Group III contains the reference haplotype (InDRE 4487) and two synonymous variants. Group IV contains the D248N mutation, seven non-synonymous and 19 synonymous variants. B) Frequency of haplotype variations in the total sample according to the major genetic groups, calculated from the number of sequencing reads for each haplotype or genetic group.

Figure 4. Genetic diversity of Influenza A(H1N1)pdm virus in the State of Morelos (2009).

The graph shows the proportion (Y axis) of each of the four genetic groups described in Table 3 in each of the 48 libraries, (X axis) ranked according to epidemiological week of the latest individual in the pool. Thus, the earliest cases are seen in the left, whereas the later cases are shown at right. The red dots represent the haplotypes of group IVb, i.e. those haplotypes that contained additional non-synonymous mutations to D248N. Black arrows in the top represent libraries in which individual amplicons were validated by Sanger sequencing. The asterisk (*) indicates those libraries which included an individual from the first wave.

Figure 5. Analysis of the relative frequency of Influenza A(H1N1)pdm virus NA mutations identified in this study (A–E) that were reported in the GenBank in 2009 (April to December) in Mexico (n  =  181) and the rest of the world (n  =  5,032).

Dotted lines represent world-wide isolates, whereas solid lines represent Mexican isolates. Red and blue lines represent alternate alleles. The 912 G/A mutation was not found neither in Morelos nor in Mexican GenBank reports, but was identified elsewhere (F). The absolute (G) and relative (H) number of isolates used for this analysis is shown for Mexico (solid black line) and world-wide isolates (dotted black line). Note that in June, July and August, only 3, 6 and 1 Mexican isolates were reported in the IVR, respectively.

Figure 6. Identification by massive sequencing of the IVb.16 (left panel) and IVb.23 (right panel) mutations associated to oseltamivir resistance.

A) View of haplotype IVb.16 and haplotype IVb.23 consensus alignment showing the frequency for each individual variant for either haplotype: A/G (black bar) and 823 C/T (red bar) for the H275Y mutation; 740 G/A (green bar) and 742 A/G (black bar) for the S247N mutation. B) Detail of aligned reads. The reference sequence (InDRE 4487) is shown in green. The mutations 823 C/T (left panel) and both 740 G/A and 742 A/G (left panel) are shown in red with a yellow background. C) Representative pyrosequencing flowgrams showing luminous intensity difference at position 823 (left panel), and 740 and 742 (right panel) with respect to the reference (InDRE 4487). D) Sanger sequencing chromatograms corresponding to individuals #791 and #281 belonging to libraries R7B4 (left panel) and R3B2 (right panel) representing the confirmation of genotypic resistance conferred by the H275N and S247N mutations, respectively.