Large-scale evolutionary surveillance of the 2009 H1N1 influenza A virus using resequencing arrays

Lee, Charlie Wah Heng; Koh, Chee Wee; Chan, Yang Sun; Aw, Pauline Poh Kim; Loh, Kuan Hon; Han, Bing Ling; Thien, Pei Ling; Nai, Geraldine Yi Wen; Hibberd, Martin L.; Wong, Christopher W.; Sung, Wing-Kin

Although we are confident that our resequencing array can successfully generate complete sequences for the H1N1(2009) virus and its variants at the current stage, we cannot rule out the possibility of reassortments between the H1N1(2009) virus and other influenza viruses. Clearly, our resequencing array cannot fully sequence such events and will generate sequences with poor quality and coverage of the reassorted segments. To investigate the effects of a reassortment event on our array, we independently amplified segments 1, 2, 3, 5, 6 and 7 of the 2009 influenza A(H1N1) virus and segment 4 of a H3N2 influenza A virus, and hybridized them onto our array. The visualization map of this experiment is shown in Figure 9 . As expected, the sequence call for segment 4 [based on PM/MM probes from the segment 4 consensus of the 2009 influenza A(H1N1) virus] is poor in quality and coverage. However, we observed that we were able to get good base calls from region 1150-1547. This region turns out to be the only significantly similar (70% matched) region between the segment 4 consensus of the 2009 influenza A(H1N1) virus and segment 4 of a H3N2 virus (CY039087). This shows that identifying regions of high similarity between the 2009 influenza A(H1N1) virus with other influenza viruses and checking if these regions have good sequence calls may be a plausible way of detecting reassortments. The drawback of this approach is that it will fail to detect reassortment of Table 3 . Hybridization intensity reduction orders found in 14 hybridization experiments certain segments where there are no regions of high similarity between the H1N1(2009) virus and the parental influenza virus. It is also difficult to annotate and differentiate every region that the H1N1(2009) virus and all other influenza viruses share similarity with. We propose an alternative approach to detect reassortments. By analysing the PM/MM hybridization intensity foldchange of high confidence calls of all eight segments, we found that the average PM/MM hybridization intensity fold-change of high confidence calls in segments 1, 2, 3, 5, 6 and 7 belonging to the 2009 influenza A(H1N1) virus is $4.5 while the average PM/MM hybridization intensity fold-change of high confidence calls in segment 4 belonging to the H3N2 influenza A virus is only 1.9. The most likely reason for this huge drop in the average PM/MM hybridization intensity fold-change of high confidence calls is that the signal gained by most of the segment 4 PM probes on our array are through cross-hybridization to the segment 4 sequence of the H3N2 influenza A virus, and thus much lower than signal gained from true specific binding. Thus, by computing and comparing the average PM/MM hybridization intensity fold-change of high confidence calls in each segment, we can identify potential reassortments in a given H1N1(2009) virus sample. Virus samples with possible reassortments can then be sequenced using . A heat map bar is used to represent the quality and coverage of its sequence calls. The locations of all mutation calls made by EvolSTAR are represented by red triangles beneath the heat map bar. Sequences with coverage <90% are automatically flagged as 'low coverage'. Other details such as coverage: percentage of base calls successfully made, match: number of base calls that match the reference sequence i.e. non-mutation base calls, strong mismatch: number of high confidence base calls that do not match the reference sequence i.e.