Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jul 1.
Published in final edited form as: Infect Genet Evol. 2019 Feb 22;71:1–6. doi: 10.1016/j.meegid.2019.02.025

Inconsistent temporal patterns of genetic variation of HCV among high-risk subjects may impact inference of transmission networks

Rebecca Rose 1, Christopher Rodriguez 1, James Jarad Dollar 1, Susanna L Lamers 1, Guido Massaccesi 2, William Osburn 2, Stuart C Ray 2, David L Thomas 2, Andrea L Cox 2, Oliver Laeyendecker 2,3
PMCID: PMC6620775  NIHMSID: NIHMS1017221  PMID: 30802530

Abstract

Hepatitis-C Virus (HCV) sequences are often used to establish networks of people who inject drugs (PWID). However, the degree to which within-host evolutionary dynamics affect those inferences has not been carefully studied. Here, we analyzed 702 longitudinally-sampled HCV E1 sequences from 88 HCV+ people who inject drugs (PWID) in the Baltimore Before and After Acute Study of Hepatitis (BBAASH) cohort. Individuals were tested for HCV RNA over multiple visits to the clinic, and the HCV E1 gene was sequenced for HCV+ samples. Genetic clustering was performed on the full set of sequences using a 3% genetic distance threshold to define epidemiological linkage. Maximum-likelihood (ML) phylogenies were inferred to assess evolutionary relationships. We found 22 clusters containing sequences sampled over five or more years (long-term clusters, LTC), of which 17 had >1 subject. In six of the multi-subject LTC, one subject had a sequence sampled >3 years earlier or later than the next-closest subject in the cluster (time-gap LTC). ML trees showed that, in three of the time-gap LTC, two subjects had identical sequences despite 7–10 years separating the sampling times. In four of the time-gap LTC for whom additional data were available, the subject with the later detected shared variant had both different variants and visits with no detectable HCV RNA (RNA-) prior to the appearance of the shared variant. In the subject with the earlier detection of the shared variant, different variants and RNA- visits were also detected in multiple cases subsequent to appearance of the shared variant. Complex patterns of shared viral variation among PWID reflect on-going re-infection, multiple transmission partners, and/or inconsistent detection of viral variants. Our results suggest that transmission events are currently underestimated by analysis of sequences at a single point in time.

Keywords: epidemiology, clustering, phylogeny, evolutionary dynamics, evolutionary rate

1. INTRODUCTION

HCV continues to be a major health care burden, despite new classes of drugs with high efficacy [1]. HCV is also still a major cause of mortality throughout the world [2]: in 2016, while up to 2 million persons were cured of HCV, 1.7 million new infections occurred [3]. The vast majority (up to 80%) of new infections in high-income countries are among people who inject drugs (PWID) [4], and an estimated 67% of PWID have been infected [5]. Especially troubling is the rising incidence of HCV infection due to opioid use in young people in the US [6].

Molecular epidemiological tools have been used to identify transmission networks among PWID [7–14]. However, molecular epidemiological analysis of HCV in PWID is complicated by several factors. The actual onset of HCV infection is rarely known. Spontaneous clearance occurs in 25–35% of infected individuals, and some individuals who spontaneously clear infection can be reinfected [15–17]. In addition, multiple viruses can establish the initial infection (coinfections) and those who develop chronic infection can acquire new viruses later (superinfection). A multitude of observational studies suggest that a substantial proportion of PWID carry mixed infections [18]. Furthermore, within-host evolutionary dynamics are complex [19–26]. Multiple co-existing lineages persist over time and are not necessarily detected at every sampling time [19, 20, 26], even using high-throughput sequencing [20, 26], likely resulting from viral population structure [19, 20, 26, 27]. Lineages may exist for many years yet still show a relative lack of evolution [19, 20, 26, 27], and the evolutionary rate is highly variable over time and among lineages [19–22, 26]. Furthermore, among-host dynamics are complicated by the possible “reversion” of host-specific substitutions upon transmission [23, 28, 29]. Thus, at any one point in time, the viruses detected may reflect a combination of reinfection, superinfection, and evolution [18–26, 28, 29].

Because these factors complicate cross-sectional analyses, we assessed longitudinal patterns of HCV in a highly-networked cohort of 88 PWID followed over ~20 years with known dates of seroconversion. We previously investigated patterns of intra-host variation in this cohort, and found that 1/3 of the subjects in this cohort had multiple distinct variants during the study period [30]. The most common pattern was that a subject had one particular viral variant early in infection, and then a second viral variant was subsequently detected at the remaining visits. Other more complex patterns included two different viral variants being detected simultaneously; the earlier variant “re-appearing” after detection of a different variant and/or and HCV RNA- visit; and a single subject having up to six viral variants [30]. In this report, we follow up on those results to investigate clustering patterns of shared viral variants over time among individuals and assess the impact on interpretation of epidemiological linkage.

2. METHODS

2.1. Ethics statement

This study was approved by the Institutional Review Board of the Johns Hopkins University (IRB# NA_00046368, Title: CD8 + T cells and the Outcome of viral hepatitis- Monthly Blood Sampling). This study of stored samples and previously collected data and the parent cohort study was conducted according to the ethical standards set forth by the institutional review board of Johns Hopkins University and the Helsinki Declaration of the World Medical Association. All participants were adults who provided written informed consent.

2.2. Subjects.

Subjects were enrolled in the “Baltimore Before and After Acute Study of Hepatitis” (BBAASH) cohort described previously, composed of PWID followed prospectively over a nearly 20 year period (1996–2016) [16]. The study followed hundreds of individuals who were frequently tested for HCV. All of the subjects either entered the study HCV seronegative, or had documented clinical evidence of a recent seroconversion. Enrolled subjects were tested and provided samples at regular intervals thereafter. For HCV RNA+ samples, sequencing was performed. The cohort included individuals who progressed to chronic infection, as well as individuals who appeared to clear infection and were subsequently re-infected.

2.3. Genetic sequence data.

For this study, we used a subset of BBAASH subjects (n=88) who had two or more sequences (range 2–31, average = 8 sequences/person) sampled over a 1–14-year period (average time span = 4.9 years) during the length of the study (1997–2016). We used E1 sequences (H77 nt 943–1288) from all 88 subjects, with a final alignment comprising 702 sequences. The generation of E1 sequence data was performed contemporaneously over two decades using methods previously described [31]. Standard procedures to prevent and recognize cross-contamination have been in place for decades in the JHU Center for Viral Hepatitis Research laboratory. These build upon the standards articulated by Kwok and Higuchi [32] and include: (i) physical separation of lab areas for blood handling, extraction/pre-amplification, post-amplification, and plasmid work, (ii) extensive use of negative controls and aerosol barrier tips, and (iii) immediate creation of multiple aliquots of each fresh sample prior to other manipulations, with segregation of aliquots for routine versus validation purposes. In addition to the precautions noted above, initial sequencing of the E1 amplicon was done within days of specimen collection, with approximately 8 positive reactions weekly.

2.4. Clustering/Phylogenetics.

We used genetic clustering to determine which sequences were likely to be epidemiologically closely related and define specific viral variants. We used HIV-TRACE to calculate the distance matrix and clusters [33]. We resolved ambiguities at 1.5% as suggested by the authors of the program [33]. In our previous study of E1-based sequence comparison [30] we used a 3.9% genetic distance threshold based on an ROC analysis which assessed the distributions of intra-patient diversity and inter-subject diversity. In this study, we used a slightly lower threshold of 3% to maximize discrimination of clusters (since clusters grow larger as genetic distance increases with this method [34]). Maximum likelihood (ML) trees were inferred using PhyML [35] in the Geneious software platform (https://www.geneious.com/) with the GTR+gamma model of nucleotide substitution.

3. RESULTS

3.1. Clusters.

Among the full dataset of 702 sequences, 93 clusters were found using a genetic distance threshold of 3%. The median duration (i.e. time between earliest and latest sampled sequences) for all clusters was 2.3 years (Figure 1). The number of clusters increased as total duration decreased (with total duration measured in one-year intervals), with a maximum of 23 clusters with <1-year total duration, to a minimum of one cluster with 5–6 years total duration. This result was consistent with the expectation that genetic distance between sequences that share a common ancestor (i.e. at time of transmission) increases with time as sequences independently evolve. Interestingly, the number of clusters in each one-year interval remained relatively consistent from 5–14 years total duration, with an average of 2.5 clusters per interval. Altogether, 22 clusters contained sequences collected across an interval of five years or more (long-term clusters, “LTC”). Of the 22 LTC, 17 were comprised of sequences from >1 subject.

Figure 1. Histogram of the time spanned (years) by the identified clusters.

Figure 1.

Open in a new tab

Each bar indicates the duration (x-axis) for 93 clusters (y-axis) found at the 3% genetic diversity threshold.

We were interested in investigating further the characteristics of these LTC. First, we plotted the sampling time for each sequence in the cluster. Surprisingly, we found that in four of the LTC containing sequences from two subjects, the earliest sample from one subject was between 7.2 and 10.7 years later than the latest sample from the other subject (Figure 2a-d). In the two LTC with multiple subjects, one subject had a sequence sampled 5.9 later (Figure 2e) and 3.5 years earlier (Figure 2f) than the next-closest sample from another subject. Collectively, these six clusters were designated “time-gap LTC”.

Figure 2. Sampling dates for subjects’ sequence in six time-gap long-term clusters (LTC).

Figure 2.

Open in a new tab

For each of the six clusters (A-F) for which a time-gap was identified, the sampling dates (x-axis) of all sequences are shown for the subjects in each cluster (subject ID on y-axis). Colors represent the sequences from a given subject. Non-overlapping sampling periods (i.e. “time-gap”) are indicated by the grey box.

3.2. Phylogenetic trees.

We then inferred ML trees for the four time-gap LTC with two subjects (Figure 3a-d) and the two time-gap LTC with >2 subjects (Figure 3e-f). In three of the time-gap LTC trees, two subjects shared identical sequences, despite being sampled 7–10 years apart (Figure 3b, d, f). In all three of these cases, the subject with the later sampled variant only had one sequence in that cluster. In one of these cases (Figure 3f), the time-gap between identical sequences (8 years) was much larger than the time-gap initially observed above in Figure 2f, because the later sampled individual (in 2011) shared an identical sequence with another subject sampled much earlier (in 2001–2003) rather than with the sequence from the next-closest sampled subject (in 2008). In two of the time-gap LTC, one sequence from the earlier sampled subject was closely related to sequences sampled 6–8 years later (Figure 3c, e). In the remaining time-gap LTC tree, sequences from each subject were grouped on separate clades (Figure 3a).

Figure 3. ML trees for six time-gap LTC.

Figure 3.

Open in a new tab

Each ML tree (A-F) contains sequences from all subjects in that cluster. Circles represent tips of the branches (sequences). Colors represent different subjects. The sampling year is noted next to the tips. Branches are scaled according to the scale bar. For the purposes of maintaining the scale across all subjects for comparison, a single ML tree was inferred using sequences from all LTC, and the clades containing the subjects from each of the six clusters are shown here.

For comparison, we also inferred ML trees for the 16 non-time-gap LTC. In two clusters, identical sequences from the same subject were sampled up to four years apart (Figure 4a, b). In one subject, nearly identical sequences were sampled 11 years apart (Figure 4b). On the other hand, some trees showed more of a staircase-like pattern of accumulating mutations as would be expected under a more constant molecular clock (Figure 4c).

Figure 4. ML trees for three representative non-time-gap LTC.

Figure 4.

Open in a new tab

Circles represent tips of the branches (sequences). Grey shades represent different subjects. The sampling year is noted next to the tips. Branches are scaled according to the scale bar. The tree was drawn in the same fashion as described in Figure 3.

3.3. History of time-gap LTC clusters.

Most subjects in the time-gap LTC had additional sequences in other clusters and/or RNA- visits, in addition to the sequence(s) in the shared cluster. A longitudinal plot of viremia and all cluster assignments (i.e. “variants”) was plotted for both subjects in each of three time-gap LTC (Figure 5a, b, c) and for the two subjects with identical sequences in one of the larger time-gap LTC (Figure 5d). Additional data was not available for one subject in remaining two time-gap LTC. The first cluster assignment represents the initial infecting virus around the time of seroconversion. In all four cases, the subject with later detection of the shared cluster previously had sequences in 1–5 other clusters and an RNA- visit (Figure 5a-d). In three cases, the subject with the earlier detection of the shared variant subsequently had a sequence in a different cluster prior to detection of the shared variant in the other subject (Figure 5a, b, d). In two cases, the subject with the earlier detection of the shared variant also had an RNA- visit prior to detection of the shared variant in the other subject (Figure 5b, d).

Figure 5. Clusters and RNA- visits for subjects in four time-gap LTC.

Figure 5.

Open in a new tab

For each cluster, the two subjects (red and blue) with non-overlapping sampling dates are shown (A-D). Time is shown on the x-axis, and the cluster designation/RNA- visit is shown on the y-axis. Cluster numbers are categorical. The bolded, italicized cluster number and grey bar indicates the LTC. NC = no cluster.

4. DISCUSSION

In this study, we investigated the temporal patterns of clusters in a cohort of PWID whose initial infection was documented and who were then carefully followed through time. Our findings underscore the complexity of interpreting HCV sequences from PWID for epidemiological purposes.

We found 22 clusters containing sequences collected across an interval of five years or more (Figure 1). In six of these clusters, the sequence sampling time for one subject did not overlap those for any other subject in that cluster (Figure 2). In fact, an identical sequence was detected in three pairs despite the two subjects being sampled 7–10 years apart (Figure 3). Furthermore, the subjects with the later shared variant had at least one different variant in another cluster and a visit with no HCV RNA prior to detection of the shared variant, pointing towards superinfection (Figure 5). In two clusters, identical sequences in the same individual were sampled over a four-year period, while nearly identical sequences were sampled over an 11-year time period (Figure 4).

There are multiple potential explanations for these patterns. Previous studies of intra-host evolution have detected extreme rate variation both within and among HCV+ subjects, as well as intermittent detection of persistent variants, consistent with a model of a highly structured within-host viral population [19–22, 26, 36]. Putative population structure in the liver and/or extra-hepatic compartments may allow distinct viral sub-populations to persist while remaining undetectable in blood [19, 20, 26, 27]. Thus, the shared variant may have been present concurrently in both subjects, but at an undetectable frequency in the blood. This explanation is consistent with studies using deep sequencing methods on blood samples which found similar patterns of alternating lineages and population structure [20, 26], and suggests that using deep sequencing would not provide additional resolution. Alternatively, the virus could have continued to accumulate mutations in the first subject subsequent to its last detection, which were lost in the new host after transmission [37]. While this could point to extremely strong selective pressure that maintains an early strain, or to a mechanism that allows for a reduced replication rate as seen for HIV [38, 39] and HBV [40], the inconsistent differences in evolutionary rates at different scales across the genome rather point to a selection and constraint mechanism [37].

These findings have major implications for the interpretation of cross-sectional HCV sequence data. Typically, two or more individuals with genetically related variants (i.e. in the same genetic cluster) are interpreted as sharing a recent transmission event. In this study, however, multiple instances of shared variants were unrelated with time. Therefore, subjects sharing identical or near-identical variants are not necessarily indicative of a recent transmission. Additionally, genetic linkage among subjects could be underestimated in studies that use only one or a few time points, even if subjects are sequenced soon after seroconversion, as infection with multiple variants could be masked by the detection of only a single dominant variant. On the other hand, the observation that genetic relationships can be detected among viruses sampled a decade or more apart supports the use of networking studies for HCV, as genetic linkage may be retained for longer periods of time than for other viruses such as HIV for which a more gradual buildup of genetic differences over time will eventually exceed the genetic distance threshold used to establish transmission groups among subjects. We note that since this cohort of 88 subjects is a small portion of the entire Baltimore PWID population, one or more links in the transmission chain are likely unrepresented. Therefore, sharing closely related variants does not necessarily imply direct transmission, particularly in PWID where high-risk behavior is prevalent among multiple individuals.

Of course, the assignment of genetic linkage is predicated on the choice of genetic distance threshold. Clear rationale for choosing distance thresholds is important for any molecular epidemiological linkage study. Previously we used a ROC-based analysis to determine the point that maximized both sensitivity and specificity based on the intra- vs. inter-subject pairwise genetic distances [30]. We used a slightly lower threshold (3%) in this study as a conservative measure to avoid over-estimating linkage among sequences, since clusters tend to exponentially grow larger as genetic distance threshold increases using a hierarchical clustering method [34]. Given that we found identical or near-identical sequences over >5-year time intervals, decreasing the threshold would not have greatly affected our results.

A possible explanation for observing the same sequence is PCR contamination. However, given the contemporaneous generation of sequence data and the quality control procedures in place this is highly unlikely to have generated the type of observations seen. The recovery of the same sequence from the same person at visits separated by months or years cannot be explained by carry-over contamination unless many other intervening specimens were also positive for this sequence, and we know from examining all E1 sequences that this was not the case.

Similar to many molecular epidemiological studies, we used bulk sequencing, which provides only one consensus sequence, thus masking genetic variation. Next-generation sequencing for full length and/or deeper sequencing may provide greater insight into the epidemiological connections within the population; however, we note the potential limitations of this approach above. On the other hand, more frequent sampling may provide greater opportunity to detect alternatively co-circulating lineages. Likely a combination of approaches will be required to fully explore which mechanism(s) account for our observations.

5. ACKNOWLEDGEMENTS

We thank the participants and staff of BBAASH.

6. FUNDING SOURCES

Support was provided by the Division of Intramural Research, NIAID. DT was funded by R37DA013806. SCR was funded by U19 AI088791. AC was funded by U19 AI088791.

REFERENCES

  • 1.Thrift AP, El-Serag HB, and Kanwal F, Global epidemiology and burden of HCV infection and HCV-related disease. Nat Rev Gastroenterol Hepatol, 2017. 14(2): p. 122–132. [DOI] [PubMed] [Google Scholar]
  • 2.Stanaway JD, et al. , The global burden of viral hepatitis from 1990 to 2013: findings from the Global Burden of Disease Study 2013. Lancet, 2016. 388(10049): p. 1081–1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Hill AM, Nath S, and Simmons B, The road to elimination of hepatitis C: analysis of cures versus new infections in 91 countries. J Virus Erad, 2017. 3(3): p. 117–123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Grebely J, et al. , Recommendations for the management of hepatitis C virus infection among people who inject drugs. Int J Drug Policy, 2015. 26(10): p. 1028–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Nelson PK, et al. , Global epidemiology of hepatitis B and hepatitis C in people who inject drugs: results of systematic reviews. Lancet, 2011. 378(9791): p. 571–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Zibbell JE, et al. , Increases in hepatitis C virus infection related to injection drug use among persons aged </=30 years - Kentucky, Tennessee, Virginia, and West Virginia, 2006–2012. MMWR Morb Mortal Wkly Rep, 2015. 64(17): p. 453–8. [PMC free article] [PubMed] [Google Scholar]
  • 7.Aitken CK, et al. , Molecular epidemiology of hepatitis C virus in a social network of injection drug users. J Infect Dis, 2004. 190(9): p. 1586–95. [DOI] [PubMed] [Google Scholar]
  • 8.Bretaña NA, et al. , Transmission of Hepatitis C Virus among Prisoners, Australia, 2005–2012. Emerg Infect Dis, 2015. 21(5): p. 765–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Cochrane A, et al. , A genetic analysis of hepatitis C virus transmission between injection drug users. J Infect Dis, 2002. 186(9): p. 1212–21. [DOI] [PubMed] [Google Scholar]
  • 10.Cunningham EB, et al. , Methamphetamine injecting is associated with phylogenetic clustering of hepatitis C virus infection among street-involved youth in Vancouver, Canada. Drug Alcohol Depend, 2015. 152: p. 272–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Jacka B, et al. , Phylogenetic clustering of hepatitis C virus among people who inject drugs in Vancouver, Canada. Hepatology, 2014. 60(5): p. 1571–1580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Matthews GV, et al. , Patterns and characteristics of hepatitis C transmission clusters among HIV-positive and HIV-negative individuals in the Australian trial in acute hepatitis C. Clin Infect Dis, 2011. 52(6): p. 803–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Pilon R, et al. , Transmission patterns of HIV and hepatitis C virus among networks of people who inject drugs. PLoS One, 2011. 6(7): p. e22245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Sacks-Davis R, et al. , Hepatitis C virus phylogenetic clustering is associated with the social-injecting network in a cohort of people who inject drugs. PLoS One, 2012. 7(10): p. e47335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Osburn WO, et al. , Spontaneous control of primary hepatitis C virus infection and immunity against persistent reinfection. Gastroenterology, 2010. 138(1): p. 315–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Cox AL, et al. , Prospective evaluation of community-acquired acute-phase hepatitis C virus infection. Clin Infect Dis, 2005. 40(7): p. 951–8. [DOI] [PubMed] [Google Scholar]
  • 17.Villano SA, et al. , Persistence of viremia and the importance of long-term follow-up after acute hepatitis C infection. Hepatology, 1999. 29(3): p. 908–14. [DOI] [PubMed] [Google Scholar]
  • 18.Cunningham EB, et al. , Mixed HCV infection and reinfection in people who inject drugs--impact on therapy. Nat Rev Gastroenterol Hepatol, 2015. 12(4): p. 218–30. [DOI] [PubMed] [Google Scholar]
  • 19.Raghwani J, et al. , Exceptional Heterogeneity in Viral Evolutionary Dynamics Characterises Chronic Hepatitis C Virus Infection. PLoS Pathog, 2016. 12(9): p. e1005894. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ho CK, et al. , Characterization of Hepatitis C Virus (HCV) Envelope Diversification from Acute to Chronic Infection within a Sexually Transmitted HCV Cluster by Using Single-Molecule, Real-Time Sequencing. J Virol, 2017. 91(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Gray RR, et al. , A new evolutionary model for hepatitis C virus chronic infection. PLoS Pathog, 2012. 8(5): p. e1002656. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Gray RR, et al. , Unexpected maintenance of hepatitis C viral diversity following liver transplantation. J Virol, 2012. 86(16): p. 8432–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ray SC, et al. , Divergent and convergent evolution after a common-source outbreak of hepatitis C virus. J Exp Med, 2005. 201(11): p. 1753–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Cox AL, et al. , Cellular immune selection with hepatitis C virus persistence in humans. J Exp Med, 2005. 201(11): p. 1741–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Timm J, et al. , CD8 epitope escape and reversion in acute HCV infection. J Exp Med, 2004. 200(12): p. 1593–604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Raghwani J, et al. , High resolution evolutionary analysis of within-host hepatitis C virus infection. J Infect Dis, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Gray R, et al. , A New Evolutionary Model for Hepatitis C Virus Chronic Infection. PLoS Pathog, 2012. 8: p. e1002656. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Gray RR, et al. , The mode and tempo of hepatitis C virus evolution within and among hosts. BMC Evol Biol, 2011. 11(1): p. 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Liu L, et al. , Spontaneous clearance of primary acute hepatitis C virus infection correlated with high initial viral RNA level and rapid HVR1 evolution. Hepatology, 2012. 55(6): p. 1684–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Rose R, et al. , Complex patterns of Hepatitis-C virus longitudinal clustering in a high-risk population. Infect Genet Evol, 2017. 58: p. 77–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Enomoto N, et al. , There are two major types of hepatitis C virus in Japan. Biochem Biophys Res Commun, 1990. 170(3): p. 1021–5. [DOI] [PubMed] [Google Scholar]
  • 32.Kwok S and Higuchi R, Avoiding false positives with PCR. Nature, 1989. 339(6221): p. 237–8. [DOI] [PubMed] [Google Scholar]
  • 33.Kosakovsky Pond SL, et al. , HIV-TRACE (Transmission Cluster Engine): a tool for large scale molecular epidemiology of HIV-1 and other rapidly evolving pathogens. Mol Biol Evol, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Rose R, et al. , Identifying Transmission Clusters with Cluster Picker and HIV-TRACE. AIDS Res Hum Retroviruses, 2017. 33(3): p. 211–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Guindon S, et al. , Estimating maximum likelihood phylogenies with PhyML. Methods Mol Biol, 2009. 537: p. 113–37. [DOI] [PubMed] [Google Scholar]
  • 36.Liu L, et al. , Acceleration of hepatitis C virus envelope evolution in humans is consistent with progressive humoral immune selection during the transition from acute to chronic infection. J Virol, 2010. 84(10): p. 5067–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Gray RR, et al. , The mode and tempo of hepatitis C virus evolution within and among hosts. BMC Evol Biol, 2011. 11: p. 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Vrancken B, et al. , The genealogical population dynamics of HIV-1 in a large transmission chain: bridging within and among host evolutionary rates. PLoS Comput Biol, 2014. 10(4): p. e1003505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Raghwani J, et al. , Evolution of HIV-1 within untreated individuals and at the population scale in Uganda. PLoS Pathog, 2018. 14(7): p. e1007167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Vrancken B, Suchard MA, and Lemey P, Accurate quantification of within- and between-host HBV evolutionary rates requires explicit transmission chain modelling. Virus Evol, 2017. 3(2): p. vex028. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES