Statistical analysis of combined substitutions in nonstructural 5A region of hepatitis C virus and interferon response.

Interferon (IFN) therapy (+/- ribavirin) is the only currently available treatment for chronic hepatitis C virus (HCV), but is not effective for a majority of patients. Several studies have correlated IFN response with substitutions in a small region of the nonstructural 5A (NS5A) gene product of HCV, termed the interferon sensitivity determining region (ISDR). Many other studies, however, have been unable to verify this correlation. To address this issue, available data from published studies was used to create a database of 675 individual ISDR sequences. The database was used to analyze substitutions in the ISDR with regard to IFN response. Combined data was analyzed by the chi-square independence test and by logistic regression analysis. Each statistical analysis demonstrated a strong correlation between IFN response and substitutions in the ISDR. A statistically significant correlation was also found between IFN-response and substitutions in ISDRs of combined studies that independently were unable to detect a correlation. The failure of these individual studies to verify a correlation seems to be at least partially due to inadequate sample size. A new model for chi-square analysis is proposed that could allow a correlation between IFN-response and ISDR sequence to be detected for data sets with less statistical power than that required for the current model. The ISDR database was also used to analyze individual substitutions in the ISDR. The results show that IFN-sensitive viruses contain a larger number and more diverse collection of substitutions than IFN-resistant viruses. Substitutions that are most likely to be tolerated or detrimental to NS5A function in the IFN-response mechanism were identified as a first step toward site directed mutagenesis of the ISDR.