Integrating Haplotype-Enhanced HIV Surveillance with National Databases Improves the Resolution of Transmission Networks, Drug Resistance, and Risk Assessment

Sanger and other consensus-based HIV surveillance methods provide limited resolution for characterizing transmission dynamics. This study evaluated the added value of integrating haplotype sequences into the DRVI pol database to improve analyses of viral diversity, transmission networks, and haplotype-level drug resistance. We deep sequenced the pol gene from 95 newly diagnosed individuals in the 2023 national survey, reconstructed haplotypes using CliqueSNV, and incorporated all haplotypes into the DRVI database for network construction. Principal component analysis and k-means clustering were used to identify subgroup structures based on viral diversity, selective pressures, and host factors, and these were linked with molecular network data. A total of 1,594 haplotypes were reconstructed, achieving a 91.1%sequence-level network inclusion rate (1452/1594 haplotypes), which was 2.4-fold higher than that of consensus sequences (37.9%, 36/95). Haplotype integration increased network connections by 2.14-fold (75 vs 35). Two subgroups were identified (p < 0.001), with high-diversity/low-selection individuals forming denser clusters (64 vs 6; p < 0.001). Among detected drug-resistant mutations, 69.7% were observed only at the haplotype level, and resistance patterns varied within the same genetic similarity-based cluster. These findings demonstrate that haplotype integration markedly enhances HIV molecular surveillance by resolving previously unrecognized links, clarifying diversity-associated network structure (which may partly reflect infection stage), and detecting low-frequency drug resistance missed by consensus methods.