(Free) random pipeline idea: nf-core/phylogenomics...
# meta-omics
j
(Free) random pipeline idea: nf-core/phylogenomics - give it an MSA FASTA, and then can run a bunch of different phylogenomics tools (NJ, MP, ML, BEAST, ...), and you could also include different substitiution models/other parameters for each of them
j
Pass it a bunch of genomes and it calls e.g. BUSCOs and builds a phylogeny with those?
👍 1
e
Givaways xD
j
For example
But more flexibly. For example in aDNA we always do it via variant/SNPs calling
So normally have a SNP table like thing, and build NJ tree from
vegan
, MP via MEGAX (although wouldn't work here, I'm sure there is an R package), ML with iqtree or RaxML), Bayesian stuff with BEAST2
d
I'm a bit skeptical of automated phylogeny, but I can see your use case. As I think phylonetworks are good for "sanity checks" of data, I'd probably include NeighborNet in the list if there's an implementation of the method that works (I always use SplitsTree, but I'm not sure we could call that from Nextflow).
👍 1
j
Automated in what sense?
d
Automated in the sense of putting all steps in a pipeline, so without inspection of intermediate results before proceeding to the next step. This is of course what GTDB-Tk does (the de novo workflow), so I'm aware that there are valid use cases.
j
I see, in my head it's: here is a premade alignment, build a bunch of trees. I would not start from reads (for example)
d
Right, a bit like GTDB-Tk but different input data. I've been on and off contemplating a pipeline for my standard procedure for protein phylogenies, as that would fit with #phyloplace, and always concluded that it wasn't a good idea.
👍 1