Understanding the evolution of HBV genome, and identifying mutations that determine phenotypic variations in different clinical and epidemiological settings are important for management of HBV drug resistance and vaccine escape. In this study, we describe for the first time the use of the EPLD-PCR approach for analyzing viral quasispecies of the HBV genome of different genotypes. WG quasispecies information obtained using this approach is derived from independent DNA molecules rather than from consensus sequence of all intra-host variants, as is the case with deep sequencing approaches 
. This EPLD-PCR novel approach provides important information for delineating the dynamics of the intra-host viral population and understanding the genetic linkage among genomic sites.
The main features of the EPLD-PCR-based approach to HBV WG quasispecies sequencing are that it: (i) is real-time-based which thereby circumvents the need for post-PCR cloning, (ii) has high sensitivity, (iii) has minimal PCR-prone errors due to the use of high-fidelity enzyme, (iv) uses the same PCR conditions for all sets of PCR primers; (v) has equal sensitivity for all six nested primers, (vi) incorporates tagged nested primers allowing use of only 2 sequence primers, and (vii) is adaptable to automation so reducing cost, hands-on time and probability of cross-contamination.
It is frequently assumed that the end-point limiting-dilution protocol for amplification of individual genetic variants depends on dilution down to a single molecule 
. Recently, we have shown that this dilution is rather determined by the probability of amplification of only a single molecule from a pool of molecules. The minimal size of the pool is defined by the efficiency of PCR amplification, with less efficient PCR conditions requiring a larger pool of molecules for single-molecule amplification 
. This finding suggests that, although amplification of long DNA molecules like HBV WG has low probability of being achieved from a single molecule of the HBV genome, a single HBV genome variant can still be amplified from a small pool of DNA molecules at that dilution. Such a principle was applied in this method developed for HBV WG quasispecies analysis.
To improve efficacy of quasispecies detection and sequencing, we devised a second-round PCR using 6 sets of PCR primers for amplification of small overlapping fragments. Theoretically, such amplification protocol may result in assembly of mosaic WG sequences if some of 6 PCR fragments were derived from different HBV variants. However, the experiments conducted in this study showed that there is a very low chance of amplifying a mixture of HBV WG at the end-point. This observation strongly validates the use of 6 PCR fragments in the second-round amplification for HBV WG quasispecies analysis. This is also proof of concept that the amplification of HBV WG at limiting dilution is narrowed down to a single molecule randomly selected from a small pool of DNA molecules, while the direct PCR amplification for consensus sequencing uses a mixture of DNA molecules. Such consideration underlies why consensus sequence-based methods are only capable of detecting substitutions present in viral quasispecies with a prevalence of >20% of the total HBV population 
Amplification of WG quasispecies through EPLD appears to have the propensity to recover both the commonly occurring genomes, and those that do not belong to the main lineage. Detection and molecular characterization of the viral population, including minor variants, at the WG level is important for understanding the development of drug resistance. For example, it is known that minority drug-resistant human immunodeficiency virus variants arising during therapy are significantly relevant clinically, as they can quickly grow out when subjected to the selective pressure exerted by the drug to then become dominant, so leading to treatment failure 
. Many published data have shown that the appearance of resistance mutations may precede the increase in viral load (virological resistance) by several months 
and could be a prognostic marker for the occurrence of viral breakthrough. However, since WG sequences of intra-host HBV variants are only infrequently available, changes in the genetic structure of the intra-host HBV populations under the selection pressure of therapeutic treatment are usually assessed using subgenomic fragments. A significant conservation of the HBV genome limits genetic variability of subgenomic regions 
, thus preventing the accurate reconstruction of intra-host HBV populations. shows changes in the genetic composition of the intra-host HBV population after the therapeutic treatment. Although application of drugs inhibiting viral replication in treated patients seems to provide the strongest selection pressure, HBV should be under many other selection pressures related to the natural course of chronic infection such as neutralizing immune responses. Interaction among different selection forces may affect the outcome of therapy and requires a thorough evaluation, which can be achieved through analysis of WG HBV sequences. This analysis is key to the identification of epistatic connectivity among genomic sites 
, and to relating this connectivity to outcomes of therapeutic treatment 
. Only a limited number of HBV mutations selected during antiviral treatment have been well characterized, and the diagnostic and public health implications of other mutations occurring at the whole genome level need further investigation.
The application of this novel technique to analysis of HBV WG quasispecies is also important for the identification and evaluation of recombination events. HBV recombination has been frequently detected during co-infection of hosts with more than one HBV genotype, presenting an additional source of variability in HBV genome 
. In the present study, we have tested HBV WG sequences using SIMPLOT for the detection of recombination and for mapping exact recombination points. Bootscan analysis of one study patient indicated that recombination emerged in the core and RT gene. It has been reported that the sites of recombination appeared different in each variant, suggesting that recombinant variants are subjected to selection, and only one or a few variants ultimately become dominant 
. The ability of EPLD-PCR analysis demonstrated here in identifying co-infection, viral sub-populations and recombination events substantiates the versatility of this approach for HBV WG quasispecies analysis.
A critical observation made in this study was that HBV-WG consensus sequences obtained by direct sequencing of PCR fragments without EPLD are not always identical to the major HBV variants in the intra-host population. In some cases, HBV consensus sequences were found to be different from the closest intra-host HBV variants at 1–2 genomic sites. This difference might be attributed to the complex pool of HBV DNA molecules present in the sample and the stochastic nature of PCR. Variation in amplification of different molecular species from the sample among 6 PCR fragments used in the protocol for consensus sequencing may result in reconstruction of hybrid WG sequences. This is overcome in the EPLD process as the amplification is limited to a single or minimal number of DNA molecules. The observation of the discrepancy between the consensus sequences and intra-host HBV variants indicates that WG consensus sequences should be judiciously used in genetic analysis. Sequence-based studies of HBV WG quasispecies should afford a more accurate assessment of HBV evolution in various clinical and epidemiological settings.
The data presented here project the strategy of amplifying full-length HBV DNA quasispecies with minimal distortion using EPLD approach. Unlike cloning approach where clonal selection is performed after PCR amplification, EPLD-PCR introduces the least PCR errors as DNA is diluted ahead of any amplification process, thereby minimizing template jumping, allelic preference, etc. There may be still concerns about primer bias and/or mismatches with at least some archived HBV sequences which might limit the outcome of their analyses. However, any improvement to the current methodology, such as using more efficient DNA polymerases (phusion DNA polymerase), can be improvised to minimize the PCR-prone errors.
In comparison to ultra-deep sequencing technologies, the EPLD approach generates fewer sequences. However, the EPLD approach readily generates HBV WG sequences. We and others have observed that assembly of such long sequences for individual variants from short reads produced by the ultra-deep sequencing technologies generates extremely challenging computational tasks 
. Additionally, the EPLD approach does not require any additional equipment besides conventional equipment for PCR and sequencing that is commonly available in many molecular laboratories, nor do the data generated require extensive computational assembly. Thus, the novel EPLD PCR-based technique described here for sequencing of individual HBV complete genome variants from a single DNA molecule is conveniently amenable to HBV WG quasispecies analysis, thereby allowing in-depth characterization of intra-host HBV populations, detection of recombination events and evaluation of epistatic connectivity along the HBV genome that facilitates investigation of genomic co-evolution in clinical and epidemiological settings.