Ran detaxizer (for phiX174) and proteinfamilies for the pilot dataset: vagkaratzas.github.io/meta-omics-sig detaxizer found minimal cases, that hadnt made it into metatdenovo's results anyway - so no need to rerun metatdenovo (198,252 proteins) and proteinfamilies (405 families) so far. Deciding next run for the pilot LMO dataset: 1. viralmetagenome: bots say NO:

1. Wrong size fraction

ERR13967264  LMO metagenome 2016-03-15 2 m depth, filter fraction: 0.2, replicate a
             WGS | METAGENOMIC | RANDOM

The LMO metagenome is material collected on a 0.2 µm filter, library_selection: RANDOM — no viral enrichment, no hybrid capture. Free virions are 20–200 nm; they pass through 0.2 µm into the filtrate, which is discarded. What you'd actually be looking at is prophage in bacterial genomes, intracellular replicating phage, and whatever NCLDV is big enough to be retained. That's a real signal, but it's not a virome, and it's not what the pipeline expects.

2. Wrong pipeline purpose — the defaults give it away

viralmetagenome is "reconstructing consensus genomes and to identify intra-host variants from metagenomic sequencing data or enriched based sequencing data like hybrid capture." Then: scaffold to centroid → iVar consensus → intra-host variants. This is the pipeline you run when you have a known-ish virus in a sample and want its consensus genome plus within-host diversity. "Intra-host variants" has no meaning for a Baltic Sea community, and RVDB/Virosaurus are the wrong references for marine phage — most contigs would simply not hit the reference pool, which is the step the whole workflow pivots on.

2. MAG: bots go YES:

1. It closes the loop on issue #100. Patch group and short_reads_platform into detaxizer's sheet, feed it to mag, and you have evidence the proposed fix is correct — not just that the current output is broken.
2. It unblocks mag → proteinfamilies, the route the metro map has always drawn and this project has never been able to test, because mag emits no protein FASTA. You'd add the ORF-calling step the table keeps flagging — and you already have run 03's 405 families from the metatdenovo route as the comparison. MAG-derived vs MT-derived proteins into the same pipeline is a genuine methods result, not just an edge tick.
3. It's the prerequisite for seqsubmit, the chain's only exit back to the archive.

1. We have viral metaGs but they aren't public yet. Won't be until they're published. (The student who's responsible have tried #C0905QCFTPE a bit.) 2. Magmap will call genes on the set of genomes it chooses. How to control what it chooses (only local, only public, a combination, only species and other options) should be well documented. Let me know if it's not.

alright, mag it is then, so I can also unlock magmap, seqsubmit etc