In addition, UDPS proved to be a powerful technique for quantitative study of the dynamics of HBV populations resulting from the multiple evolutionary Mechlorethamine hydrochloride pressures of sequential NA therapy. The capability to clonally analyze thousands of sequences disclosed combinations of aa 3,4,5-Trimethoxyphenylacetic acid substitutions occurring in the same genome during antiviral treatment. Identification of mutations in extremely low percentages with an acceptable degree of confidence is limited by the number of independent template molecules obtained from the sample analyzed, the coverage or number of reads obtained per base, and the number of artifactual errors generated during PCR amplification and pyrosequencing. For these reasons, all the samples selected for the current study carried a high HBV viral load, and the high-fidelity DNA polymerase Pfu turbo linked to a Poisson-based computational algorithm was used to bypass the artifactual errors. With this approach, variants comprising as little as 0.03% of the HBV quasispecies could be detected. UDPS analysis showed that aa changes known to be associated with NA resistance above this detection limit were present in low percentages in the baseline HBV quasispecies, probably representing a background due to the natural dynamics of the viral quasispecies. Moreover, variant combinations were not detected in the baseline quasispecies. These results suggest that if the resistance changes were present, they would be in percentages below the detection limit. Therefore, higher sensitivity than is reported here seems to be required for detecting combined variants at baseline. In the sequentially treated patient, the relative frequency of NA-resistant substitutions in the baseline quasispecies did not seem to be predictive of subsequent LMV treatment outcome or RT variants selected at VBK. The high sensitivity of UDPS also made possible identification of conserved residues. The residue rtL155 was found to be the most highly conserved at both pre-treatment and throughout sequential treatment, a previously unreported finding. According to a structural analysis based on a 3Dmodel of HBV RT, this leucine residue is located at the external surface of RT, and has high hydrophobicity, features that suggest a role in protein-to-protein interactions. The low overall frequency of aa substitutions in rtL155 was even lower than that observed in residues deemed essential for viral polymerase function, such as rtY203, rtD205 and rtD206 from the conserved YMDD catalytic motif. Considering the essentiality of rtD205 and rtD206, which are part of the catalytically essential aspartic acid triad of HBV RT, a higher than expected variability was found. In this sense, it has been reported that substitutions in rtD205 result in replication-defective HBV variants that can be trans-complemented in vitro by wild-type polymerases. This mechanism 
may explain the finding of genomes with aa substitutions in rtD205, which may replicate through trans-complementation with a helper wild-type HBV polymerase in the same hepatocyte. The HBV genome has an extremely overlapping structure. With UDPS, we were also able to study epitopic regions of the S ORF that overlap the RT region. In the baseline viral populations, the S ORF showed significant percentages of substitutions that lead to a stop codon. These mainly overlapped RT positions related to NA resistance variants, and some of them delete important envelope residues involved in viral infectivity and/or possible interactions with core proteins. The envelope stop codon variant sW172*, which is related to the major NA-resistant variant rtA181T.