Generating human papillomavirus (HPV) reference databases to maximize genomic mapping

Area of Science:

• Genomics
• Bioinformatics
• Virology

Background:

• Genomic analysis of human papillomaviruses (HPVs) requires accurate reference databases for read mapping.
• Existing HPV sequence databases like Papillomavirus Episteme (PaVE) are often gene-specific (L1 gene) and use complex alignments, complicating reference database generation.
• This leads to the creation of custom, per-analysis databases, which is inefficient.

Purpose of the Study:

• To develop a de novo strategy for generating a comprehensive and representative HPV reference database.
• To improve the accuracy and efficiency of HPV read mapping in genomic experiments.
• To provide a standardized, high-quality reference set for HPV genomics research.

Main Methods:

• A de novo strategy was employed, utilizing all human papillomavirus (HPV) sequences from the NCBI database.
• Oligonucleotide frequency profiling of entire HPV sequences was performed.
• Hierarchical clustering was applied to group and select representative sequences for the reference database.

Main Results:

• The proposed strategy successfully identified a subset of highly representative HPV sequences.
• The generated reference database demonstrated suitability for mapping the majority of mappable reads unambiguously in HPV capture experiments.
• The method offers a more robust and less complex approach compared to existing database generation methods.

Conclusions:

• The oligonucleotide frequency profiling and hierarchical clustering strategy provides an effective method for constructing HPV reference genomes.
• This approach enhances the accuracy of HPV read mapping, reducing ambiguity and improving genomic analysis.
• Recommendations are provided to further optimize HPV mapping, and the generated FASTA files are publicly available.