Comparison of biochip-based genotyping with NS5B sequencing. Table Table11 shows the results obtained by hybridization selleck products on the biochip and direct sequencing of NS5B segments. All (100%) of the 345 HCV RNA-positive sera analyzed were successfully genotyped by biochip hybridization. They included samples infected with all six HCV genotypes. TABLE 1. Comparison of HCV genotyping obtained by NS5B sequencing with that obtained by hybridization on the biochip The samples included subtypes 1a, 1b, 1d, 1e, 2a, 2b, 2c, 2i, 2j, 2k, 2l, 3a, 4a, 4c, 4d, 4f, 4h, 4k, 4p, 4r, and 5a and samples of undetermined subtypes of genotypes 1, 2, 4, and 6, as determined by sequencing. The two methods were concordant for the subtypes of 329/330 samples (99.7%), with a kappa coefficient of 0.996 (P < 0.00001).

One sample identified as 2c by NS5B sequencing was identified as 2k by NS5B biochip analysis. The NS5B sequencing method failed to determine the subtypes in 8 samples (2.3%), and the NS5B biochip methods failed in 12 samples (3.5%) (P = 0.36). Samples with an undetermined subtype by NS5B sequencing were identified as 1a, 1b, 2k, 4h, and 4r by NS5B biochip analysis. Samples with an undetermined subtype by NS5B biochip analysis were assigned to subtypes 1d, 2a, 2j, and 2l by NS5B sequencing. The subtypes of 5 samples were not determined by either method. DISCUSSION The gold standard for HCV genotyping remains PCR amplification followed by sequencing of one of the phylogenetically informative coding regions of the HCV genome, such as NS5B or core/E1, and comparison with the consensus sequences in GenBank or the Los Alamos hepatitis C virus databases (14).

We have developed a novel microarray-based assay for identifying the HCV genotype and subtype and evaluated it in comparison with the phylogenetic analysis of the NS5B region as a reference method. The new NS5B microarray assay and NS5B sequencing were in almost complete agreement. The assay relies on hybridization of a 380-nt NS5B fragment with oligonucleotides specific for HCV genotypes and subtypes immobilized on a biochip. The reliable identification of each individual genotype and subtype required the design of several oligonucleotides for each of them, in consequence of the variability of the NS5B region.

The results were interpreted using an original algorithm that included preliminary processing of the hybridization signal intensities from the biochip elements and comparison of signals from elements within the sets of genotype-specific probes and then from sets of subtype-specific probes. The new method enabled us to determine all six HCV genotypes with a sensitivity of approximately 2.0 �� 102 IU/ml of HCV RNA. This analytical performance using biochip-based genotyping and subtyping is comparable to that of commercially available assays (50), including the new generation of line probe assays (49). The new method was tested on 345 HCV-positive GSK-3 samples.