Host prediction for viral MAGs.

[Jiulong Zhao]

Dear Prof. Roux:

‌ Thanks for your effort in developing such a strong viral-host-prediction tool! Recently, I tried this user-friendly tool on my viral dataset and the results were clear and solid.

‌ Most recently, PHAMB and vRhyme (or other tools) were developed to retrieve or bin the viral MAGs (vMAGs). So, here, I want to know if the iPHoP tool can predict hosts of vMAGs. Or if I can concatenate the viral sequences from the vMAGs with “NNNN“ in gaps before running the iPHoP prediction?

‌ Thank you so much! Looking forward to your reply!

Jiulong

‌

[Simon Roux]

Dear Jiulong,

This is a good point. At this moment, there is no specific support for vMAGs in iPHoP. I think however that your proposition (transforming vMAGs into scaffolds by adding “NNN” in gaps) should be ok. Note that by default, iPHoP will discard any sequence with more than 10% of Ns. If this causes any issue, you can disable this pre-processing by using the argument “--no_qc”.

‌

Best,

Simon

[Jiulong Zhao]

Dear Prof. Roux, Thanks for your reply! I will try this as you suggested! By the way, I also want to know if the host taxonomies were assigned according to the GTDB taxonomy. Do you think the GTDB taxonomy is more acceptable or accurate than NCBI taxonomy?

Thank you so much!

Jiulong

Jiulong Zhao QIBEBT - PhD student |

[Simon Roux]

No, we don’t necessarily believe the GTDB taxonomy is more acceptable or accurate, however GTDB does come with a phylogenetic tree framework, i.e. it is easy to get the phylogenetic distance between two genomes from the GTDB database. Since we use this distance in iPHoP, using the GTDB database as a whole and the associated GTDB taxonomy made sense.

[Jiulong Zhao]

Dear Prof. Roux, Thanks for your kindly reply! Best wishes! Jiulong Jiulong Zhao QIBEBT - PhD student |

[Simon Roux]

Resolving for now - adding "Accepting vMAGs as input" to the todolist