A total of 119 influenza viruses were isolated from 98 individuals in 78 households (pandemic H1N1 (pH1N1)= 61; seasonal H1N1 (sH1N1)= 4; seasonal H3N2 (sH3N2)= 54). Of these 119 viral isolates, 42 were sampled from 21 individuals (pH1N1= 14; sH3N2= 7) who were positive in two consecutive visits. Twenty pH1N1 and 15 sH3N2 viruses were isolated from households with confirmed secondary transmission (pH1N1=8; sH3N2=8). On one occasion, a pH1N1 and a sH3N2 viruses were isolated from different members of the same household. The phylogenetic trees of HA and NA sequences of each subtype are shown (–;
Supplementary Figs. 1–
3). As the HA phylogeny were found to be more informative than that derived from the NA gene, all the subsequent phylogenetic analyses are based on the HA sequences. Four sH1N1 viruses (A/Brisbane/59/2007-like) were detected only once in different households and they were not studied further (
Supplementary Fig. 3). All viral sequences reported in this study have been deposited in GenBank (Accession numbers JN256681-JN256918).
Global genetic sequence analysis of pH1N1 has revealed the presence of at least 7 distinct clades of virus.
8 Phylogenetic analysis revealed that clades 3, 6 and 7 viruses were detected in this study (). None of these viruses were found to be genetically similar to the first laboratory confirmed case in Hong Kong (
Supplementary Fig. 4, yellow dots).
6 Similar to findings concluded in some Asian countries
8, 9 the majority of our isolates belong to clade 7. These results also indicated that there were multiple introductions of pH1N1 viruses into Hong Kong. To further illustrate this point, all the publicly available pH1N1 sequences sampled before this study were included for further analysis. Many of our clade 7 viruses were found to be clustered with those isolated from other areas (
Supplementary Fig. 4, red dots). Interestingly, some of our clade 7 viruses formed a unique branch in the tree (
Supplementary Fig. 4, green), suggesting that this cluster was associated with a single introduction. A dating analysis on these viruses was performed
10 and this particular subclade was estimated to have been introduced into Hong Kong on the 3
rd June (95% CI: 11
th May to 25
th June;
Supplementary Fig. 5, highlighted in green). This agreed with our previous estimate that community transmission of pandemic H1N1 in Hong Kong began in early June.
11For the sH3N2 isolates, 3 viral lineages were detected in 39 households (A/Brisbane/10/2007-like =1, A/Victoria/208/2009-like =5 and A/Perth/16/2009-like =33; ).
12 Using other sH3N2 viruses sampled from 1
st January 2008 to 20
th June 209 as references, multiple introductions of A/Perth/16/2009-like viruses into Hong Kong were detected (
Supplementary Fig. 6). In particular, a large number of our A/Perth/16/2009-like viruses were found to be associated with a single introduction (
Supplementary Fig. 6, highlighted in green). Subsequent dating analysis revealed that this introduction occurred around the 19
th April 2009 (95% CI: 24
th Feb 2009 to 25
th May, 2009).
Previous studies on sequence variation in human influenza virus primarily relied on studying viral sequences available from public databases or viruses sampled at a community level with little accompanying epidemiological data. By contrast, there is little information on viral sequence variation between host generations/successive transmission events.
13, 14 Gubareva et al studied the HA sequence variation during seasonal influenza virus transmission in 19 informative families.
14 In their study, no viral sequence variation within an individual/family was found, suggesting that there is a substantial genetic conservation during virus transmission within households. Our data indicated that the average sequence identity between viruses sampled from the same individual/household was high and was much greater than the one deduced from viruses that had no direct epidemiological links (; p<0.001). In addition, we found that the intra-household and intra-host sequence variations were not statistically significantly different from each other (p>0.05). In contrast with previous findings,
14 we found evidence of intra-household or intra-host sequence variations in both HA and NA genes of pH1N1 and sH3N2 viruses. Two of these nucleotide variations (N=6) resulted in non-silent mutations (
Supplementary Table 1). These mutations were also detected in subsequent confirmed secondary infections of household members and/or confirmed by repeated sequencing.
| Table 1Average sequence identity of viruses sampled within an individual, a household and a community. |
Using the intra-host HA sequence variations and sampling dates as references (), the nucleotide mutation rates in the pandemic H1 and seasonal H3 genes were estimated to be 1.032 x10
-2 and 2.052 x10
-2 nucleotide substitutions/site per year, respectively. These data agree with previous findings that pH1N1 and sH3N2 have comparable evolutionary rates,
8, 15, 16 but our estimates were about 2.6–4.5 times higher than those estimated from viruses that were sampled at a global level.
15, 16 This might be partly explained by the hypothesis that not of all these intra-host variants are fit for sustainable transmission at a community level. As the sample size of specimens from the same host is small, the implication of intra-host sequence variation on the evolution of human influenza virus might require further large-scale analyses on intra- and inter-host viral samples. Nonetheless, our data suggested that both pH1N1 and sH3N2 have similar intrinsic properties to generate infectious variants.
The publisher's final edited version of this article is available at 
