Hi all,
I am a total beginner to qiime2 with my very first set of real 16s sequences to analyze! 
I am ready to build my tree using:
(qiime) kmorris@atacama:/Net/Groups/BGI/people/kmorris/for_qiime2$ qiime phylogeny align-to-tree-mafft-fasttree \
--i-sequences rep-seqs-nonchimeric-wo-borderline.qza
--p-n-threads 24
--o-alignment fast/aligned-rep-seqs.qza
--o-masked-alignment fast/masked-aligned-rep-seqs.qza
--o-tree fast/unrooted-tree.qza
--o-rooted-tree fast/rooted-tree.qza
However, I ran into the same issue as some others:
Plugin error from phylogeny:
Command '['mafft', '--preservecase', '--inputorder', '--thread', '24', '/scratch/tmp/qiime2-archive-uo73mmg3/d93f2fd1-d498-4da4-ad93-7bd0eb851cfc/data/dna-sequences.fasta']' returned non-zero exit status 1.
Because I am a beginner, I thought it would be worth checking that the input file is in good shape, before running again with more RAM (last run was on a computer with 250 GB).
Here is a link to my qzv: table-nonchimeric-wo-borderline.qzv - Google Drive
Is this just too many sequences for the align-to-tree-mafft-fasttree method?
Is there a way to calculate the memory needs for running?
Many thanks is advance for your support. It took a while to make this data happen, so I am very happy to be here. Let me know what you think.
Hi @Kendal_Morris,
Could you please report the full error message? (re-run your command with the --verbose flag to print the complete error message to your terminal). I am not sure that you have the same exact underlying error as that other post — the "exit status 1" is just indicating that the subprocess failed
No, your dataset does not look extraordinarily large. This pipeline should be able to handle that.
We have not benchmarked this, though the mafft developers may have a better idea about this. I suspect the answer is no.
Maybe the different error is coming from the fact that the code was killed, rather than failing on it's own. (Wasn't my idea, may it would have finished on it's own).
However, since then I successfully aligned my sequences using qiime alignment mafft --p-parttree and the job finished in just a few hours (250 GB RAM, 24 threads).
Currently qiime phylogeny fasttree is running on the same machine (18 hours and counting). I'm hopeful it will finish soon.
Glad to hear it is working so far! align-to-tree-mafft-fasttree is a pipeline of multiple commands so the error you had before could have arisen at any point. Please use the --verbose flag or open any log files generated with an error message if another does occur. Fingers crossed everything runs smoothly from here. 