Abstract
Genome analysis of hepatitis B virus (HBV) in patient sera is helpful for monitoring treatment. We developed an improved version of a DNA microarray to identify HBV genotypes and to detect mutations of interest in the S, Pol, Core, and X genes. It includes an automated software analysis of fluorescence values for simpler, more robust data interpretation. In this version, probes were added to identify genotype H, to analyze 155 additional positions, and to detect 561 additional polymorphisms. Sequences were added to the alignments to resolve hybridization problems due to natural polymorphisms in the vicinity of important codons. The duplex PCR protocol allowed whole-genome analysis in a single tube. An alternative nested-PCR protocol allowed genotyping and mutations in S and reverse transcriptase (rt) genes in patients with low viral loads, as demonstrated in patients with less than 400 HBV genome copies/ml. Reproducibility was high, with variation coefficients lower than 3%. Only 0.57% of 20,771 codons from 253 samples could not be identified. The concordance with Sanger sequencing for the identification of codons improved from 92.8% to 95.7% with the improved version. Concordance was higher than 91% for codons associated with resistance to lamivudine, emtricitabine, telbivudine, famciclovir, entecavir, and tenofovir with vaccine escape and for pre-Core mutants. Concordance was lower for adefovir resistance mutations (68.6%) and mutations in the basal core promoter (60.3%), probably because hybridization efficiency was affected by the low GC content of the probes. A concordance of 93.7% with sequencing for genotype identification was observed in 190 specimens, lower than that obtained with the first version, possibly due to mixed virus populations.