Hello, I was wondering if it has ever been discuss the development of a pipeline focused on identifying mobile genetic elements within metagenomes or MAGs?

I am not aware of a meta-omics related pipeline that does that, and I can't see anything related when searching the channel

In the mag pipeline they added genomad to identify the viruses. I know that genomad, identifies plasmids as well but I dont know if it does both automatically or it has to be selected.

we got this in my lab: github.com/EBI-Metagenomics/mobilome-annotation-pipeline which has nf-core modules that overlap funcscan, and a couple of viralmetagenome

I've been told that MOB-suite is the best tool to find plasmids, have you use it? Do you agree or is genomad better?

not my field of expertise I am afraid 😅 other people are developing that

This looks great but the way! Thanks for sharing!

At what granularity would you like to find them: reads, contigs, MAGs or something else? If it's MAGs I suppose you could just run the existing tool, if it's contigs funcscan would be where to implement it -- I guess. Anything not in the main chromosome tends to be lost from MAGs of course, and these elements might be enriched in plasmids, speaking for contigs (possibly coming out of the mag or metatdenovo pipelines) being best

I initially thought contigs because MAGs tend to not have many MGE but if they are made with long reads sequences maybe they might no?

I tihnk funcscan would maybe make sense, for MAGs/genomes/contigs, but It hink I've had discussions with @Alexandru Mizeranschi and/or @Olga Brovkina about this (but I can't remember what I said at the time 🙈 )

Yes, I was not aware of that pipeline. That is pretty much what I was thinking about

We talked about genomad , and you mentioned that one disadvantage is that it doesn’t really have an analogous/complementary tool. As I understood it, for a robust pipeline it’s preferable to have at least two similar tools so that their results can be cross-checked against each other.

genomad searches for plasmids and viruses and I dont know if there are others than do both as well but mob-suite is suppose to be good for plasmids and for just viruses there are others as well. I just read a about a new tool called DeepMobilome and it looks promising but I havent done my due diligence on it yet

Have you worked with ISESscan? github.com/xiezhq/ISEScan

No, I actually havent work with any of this tools yet. I am starting a new postdoc that is about mobile elements and my first instinct was to go on nfcore to see if there was a pipeline design for them. Havo you work with it? do you recomended it? I am collecting all the tools that I think can be useful to then decide what to use

Ahhh thanks for the reminder @Olga Brovkina - exactly, there were multiple types of MGEs and funcscan is more focused on comparing on more specific things

Yep, I estimated the safety of bacteria based on three tiers: 1. proximity of ARG to the conjugation system and insertions, 2. genetic context of ARG- chomosomal/plasmid/mob. element; 3. clinical significance of ARG. So, would be great to automate it