Replidec - Use naive Bayes classifier to identify virus lifecycle from metagenomics data

DB update:The updated RepliDec training database comprises 926 curated phage genomes from the PHACTS database and verified RefSeq records. Temperate and virulent phages were balanced at a 1:1 ratio, while duplicate genomes and sequences sharing ≥95% ANI with the Mavrich test dataset were removed to minimize data leakage. These genomes yielded 39,016 predicted proteins, which were clustered to construct the protein-feature database used by the Naive Bayes classifier. Overall, the updated database is smaller but more balanced, reliable, and suitable for independent performance evaluation.