HoPhage

HoPhage identifies the host genus of phage fragments in metagenomic and metaviromic datasets to support phage–host interaction and microbial ecology studies.

Key Features:

• Dual-Module Integration: HoPhage employs a dual-module approach combining deep learning algorithms with Markov chain models to improve host prediction from phage fragments.
• Performance on Short Fragments: Performs well on short phage fragments prevalent in metagenomic and metaviromic data.
• Wide Candidate Host Range: Handles broad candidate host ranges and diverse taxonomic compositions, producing predictions at the genus level.

Scientific Applications:

• Phage host identification: Enables identification of host organisms from metavirome and metagenomic data at the genus level.
• Phage–host interaction studies: Supports investigation of interactions between phages and bacteria or archaea that influence microbial community dynamics.
• Ecosystem and biogeochemical studies: Facilitates studies of phage impacts on nutrient cycling and ecosystem functions.

Methodology:

HoPhage combines deep learning algorithms to capture complex sequence patterns with Markov chain models to probabilistically infer the most likely host from sequence characteristics and known phage–host relationships.