Transmission of Hepatitis C Virus among Prisoners, Australia, 2005–2012

Ongoing transmission is associated with drug injection.

cohort. In order to do this, the evolution between pairs of sequences as shown by pairwise patristic distances over time in longitudinal samples was considered.
5. Identify a clustering threshold by implementing a search algorithm starting from the identified minimum percentile threshold value and increasing its value by 0.001. For each incremental step, consider the median pairwise patristic distance of each between-host cluster detected. If all the median patristic distances of the detected between-host clusters are less than or equal to the cut-off pairwise patristic distance, then increase the percentile threshold and identify the new set of between-host clusters. If the median pairwise patristic distance of any of the clusters is above the cut-off pairwise patristic distance then regard the previous threshold (current threshold -0.001) as the optimal clustering threshold.
6. Identify between-host clusters detected using the optimal clustering threshold (identified in 5) as likely clusters of recent HCV transmission.
The pairwise patristic distances that allowed detection of between-host clusters was examined starting at 0.001 and ending at 0.48 for gt1 and 0.5 for gt3, respectively. The lower value was defined as the minimum percentile threshold where only within-host clusters were detected, while the upper values were defined as the maximum percentile thresholds where all sequences were included in a single between-host cluster ( Figure 2 in main article text). The optimal cut-off patristic distance representing recent transmission clusters was determined firstly by consideration of longitudinally collected within-host sequences representing a measure of the rate of within-host diversification of HCV genomes. The maximum pairwise patristic distances calculated among within-host sequences was 0.099 for gt1 and 0.095 gt3; hence these values were utilised as the cut-off for designation of between-host clusters.

1: Analysis of a single source outbreak of HCV transmission.
The evolution of genetic diversity that arises over time from a single source outbreak was also examined. To do this, publicly available consensus sequences from a cohort of Irish women (n=10) infected with gt1b HCV from a single donor via blood transfusion was utilised (3). One consensus sequence was obtained from the source in 1977, and one consensus sequence was obtained from each of the ten recipients at two later timepoints, in 1996 and 2000. A phylogenetic tree was generated using the E1-HVR1 sequences including both the infected recipients and the source. The pairwise patristic distances was measured between all sequences from the resulting tree and portrayed in relation to the time interval between the sampling time points.
Pairwise patristic distances between Core-NS3 (thus including E1-HVR1) sequences These values indicate that pairwise patristic distances in the Irish cohort were much less than the cut-off pairwise patristic distances identified for gt1 and gt3 in the analysis cohort. It should be noted however, that two between-host sequence pairs (subjects 117 and 461 from transmission Cluster A, and subjects 304 and 357 from transmission Cluster B) exhibited similar patristic distances to those found among the Irish cohort samples.

2: Analysis of rare variants from within-host viral quasispecies
The amount of genetic diversity within the quasispecies of a single subject was To address the extent of viral diversity within a single host and the impact of transmission of a 'diverse' minor variant, the distribution of pairwise patristic distances between E1-HVR1 variants within the quasispecies in samples collected at two timepoints within one year post-infection from two subjects with primary HCV infections which became chronic (subject 023 and 240). Deep sequencing data were available for these two subjects (4), with Page 6 of 10 frequencies as low as 1% in the viral population. Using these data, separate phylogenetic trees were constructed for the two subjects (Technical Appendix Figure 2 A

Examination of between-host clusters detected above the optimal cut-off
Two more clusters were detected just above the selected patristic distance thresholds. A putative between-host cluster was detected in the gt1 data, containing sequences from subjects 247 and 418 (designated as Cluster D, Figure 1 in main article text), with a median pairwise patristic distance of 0.149 (0.05 above the cut-off). Similarly in the gt3 data, a putative betweenhost cluster was detected consisting of sequences from subjects 089 and 082 (designated as Cluster E, Figure 1 in main article text) with a mean patristic distance of 0.246 (0.051 above the cut-off). No prison co-location episodes were found for the two subjects in Cluster D (Technical Appendix Table 1). For putative cluster E, subject 082 was identified as a possible source of transmission and was estimated to have become viremic with gt3 on June 09, 2006. Subject 082 was co-located with subject 089 in a prison for 14 days (August 27 until September 4, 2006), but denied injecting and sharing of injecting equipment during the period of co-location, while subject 089 reported otherwise. An estimated 12 months after co-locating with subject 082, subject 089 was found to be viremic with HCV gt3.