Is there a problem with the low number of high-quality phages detected?

[baidefeng]

Hi,
I tested this method using one sample of paired-ended metagenome sequencing data of human microbiome. A single-ended decompressed file is approximately 10GB. And the final final.fasta file only contains 13 sequences. Based on previous knowledge, there are a great many gut phages. Is this result normal?

[panguangze]

Please try lowering MIN_LEN in the config and delete the 05-furth and final_results folders rerun. By the way this may introduce false positives and fragmentation.

[baidefeng]

Please try lowering MIN_LEN in the config and delete the 05-furth and final_results folders rerun. By the way this may introduce false positives and fragmentation.

Thank you for your reply. So MIN_LEN = 10000 is the optimal selection to minimize false positives and fragmentations? In the CRC research case study in your paper, approximately how many phage sequences can be obtained per sample? The 2153 phages from FengQ_2025 and 1661 phages from YuJ_2025 are all unique phages? The 5306 high-quality phage genomes means that a unique phage correspond to multiple phage genomes? In the final.txt file, some lines have one phage name and some lines have multiple phage names. Can I understand that each name in each line represents a genome? How should I map these names to the sequences in final.fasta? And sorry for the lengthy questions and the time they take. Thank you.

[panguangze]

Yes,  MIN_LEN = 10000  is a conservative selection designed to prioritize precision.
In the FengQ and YuJ datasets, we recovered an average of approximately 14 phages per sample.
We performed individual assembly for each sample without cross-sample clustering. they are rarely 100% identical across samples. Since phages of the same species might exhibit structural diversity. Thus, we preserved specific variants rather than merging them according the ANI percent similarity.

On File Format: Each line in final.txt  corresponds to a single record in final.fasta