Meanwhile I played a bit with Clawbio skills

lit-synthesizer

and

ncbi-datasets

and added some more seemingly interesting datasets (7 atm) for our case in the shared doc (see paragraphs

Dataset Decision Matrix

and

Dataset Candidates

): https://docs.google.com/document/d/1wpd5-Ar8zu7xohKP2PzyNoCwNGRgQeDNqnNrC2KDBnw/edit?tab=t.0 Could you all please have a look at these 7 and vote below for the best case for a chained analysis with our meta'omics pipelines, before the next meeting? According to the bots:

The Herold wastewater dataset (PRJNA230567) is now confirmed as the only public dataset with all 4 omics layers (environmental). No host-related dataset with 4-omics coverage was found. The recommended path surfaced by the search: PRJNA230567 as primary benchmark (max pipeline coverage) + PRJNA289586 as host-related companion (gut/T1DM narrative).

Just as comment: PRJNA682552 is the only dataset with Nanopore shotgun metagenomics reads, all others seem to have only Illumina? But thats not that important because we just want maximal pipeline coverage. I think the Dataset Decision Matrix is 1 point too small for the first two datasets (i.e. the weighted total is 1 too less, maybe my calculation is off?)?

you are right, bots can't do math xD

Also what are all the acroyyms?

yeye, these 3 and amplicon

KS is amplicon?

no, that's just from the name of the study

Such info normally would go at the top BEFORe you read it 😉

Proposed pipeline chain(s) to be discussed further:

Selecting genomes for building a suitable wastewater database with createtaxdb may be a bit harder

feel free to drop any comment on the doc 😄 https://docs.google.com/document/d/1wpd5-Ar8zu7xohKP2PzyNoCwNGRgQeDNqnNrC2KDBnw/edit?tab=t.0#heading=h.e5cov2afhzqp

Sorry yes indeed

Reminder to vote on the dataset poll; maybe you all can take a look into these over a chill weekend coffee :D

I had a look yesterday, but only made comments in the doc so far. Since we don't have any proteomics tools in the SIG, and not that many in nf-core at all, I was wondering a bit how to deal with that. Just ignore, or try to get some tools? We're currently working on a project in which we have metaG, metaT and metaP and making comparisons is very difficult because the proportion of ORFs discovered in the latter is very small, which is maybe another reason to skip metaP from this project. I'm not sure if setting the bar at amplicon + metaG + metaT would change so much though, but that's where I got stuck.

There are some pipelines in early stage discussions for proteomics, but I agree that it might be better to completely disregard those, at least for now. amplicon + metaG + metaT seems the most reasonable!

This probably increases the number of possible datasets quite a bit. If any of the ones we have still suit us, I suppose we could just go ahead and vote, but if there are actually better datasets that only contain the three, maybe we should have look? To be precise, only if we see problems with the current selection.

OFc, everybody is welcome to add more datasets to the pool 😄

As I tried to say, I'm not after adding more -- the seven are difficult enough to evaluate -- but I got the impression that none of them was what we were looking for.