FusionPHI_data

We used 672 phage-host interaction (PHI) pairs provided by the PHP study were used as the baseline in this study. The ratio of positive samples to negative samples was 1:1.Since the PHP dataset only provides the accession number and taxonomy information of PHI pairs, and does not directly contain specific genome sequences, the genome characteristics of phages and hosts need to be extracted independently for each PHI pair. Genomic features of phage and host were extracted for each PHI pair. For the host genome, the scientific nomenclature of bacterial species was extracted from 672 PHI pairs, and 82 target species were obtained after de-emphasis, and ORF sequences were obtained from NCBI; for the phage genome, 333 phage accession numbers that can be used for direct querying were extracted from 672 PHI pairs and ORF sequences were obtained through them directly. Finally, a dataset containing 82 bacterial as well as 333 phage genomes was constructed for model training and validating.