Fragment-insertion sepp error: non-zero exit status

[aravenscraft (Alison Ravenscraft)]

Hi!

I am trying to run qiime fragment insertion with SEPP. I'm using a custom SEPP database I built myself following @dethlefs' very helpful instructions here: Trouble generating SeppReferenceDatabase qza - #2 by dethlefs.

I generated the SEPP database using qiime tools import --type SeppReferenceDatabase, and now I'm trying to test whether it works by running fragment insertion on my first five ASVs. Here are my reference database and the 5 SV sequences:
SeppRefDB.qza (16.6 KB)
sv1sv5.qza (6.1 KB)

I ran:

qiime fragment-insertion sepp --i-representative-sequences sv1sv5.qza --i-reference-database SeppRefDB.qza --o-tree insertion-tree.qza --o-placements insertion-placements.qza

But got the error:

Plugin error from fragment-insertion:

Command '['run-sepp.sh', '/var/folders/1m/v2hn0_5s66nf5z894vbq8hq40000gq/T/qiime2/alison.ravenscraft/data/5f7b4326-9c05-40ba-9d8e-b2fa7dd0adc5/data/dna-sequences.fasta', 'q2-fragment-insertion', '-x', '1', '-A', '1000', '-P', '5000', '-a', '/var/folders/1m/v2hn0_5s66nf5z894vbq8hq40000gq/T/qiime2/alison.ravenscraft/data/b376f2a2-d9a4-4f3d-b392-7a4229f3bcf8/data/aligned-dna-sequences.fasta', '-t', '/var/folders/1m/v2hn0_5s66nf5z894vbq8hq40000gq/T/qiime2/alison.ravenscraft/data/b376f2a2-d9a4-4f3d-b392-7a4229f3bcf8/data/tree.nwk', '-r', '/var/folders/1m/v2hn0_5s66nf5z894vbq8hq40000gq/T/qiime2/alison.ravenscraft/data/b376f2a2-d9a4-4f3d-b392-7a4229f3bcf8/data/raxml-info.txt']' returned non-zero exit status 1.

Following @thermokarst's advice here, I re-ran the command with the --p-debug and --verbose options. Here is the output:
sepp debugging.txt (7.7 KB)

A few previous posts have suggested this error is caused by a memory issue, but my reference phylogeny only has 122 leaves, and I'm only trying to place 5 ASVs onto it. Furthermore, I followed @dethlefs and @Stefan's suggestion here to trim the reference alignment to just the fragment region being placed, so my reference alignment (as well as my 5 ASVs) is only 468 bp long. I have also tried running fragment insertion via QIIME2 in Galaxy - that also failed. So it seems like this shouldn't be a memory problem - but maybe I'm wrong.

Any suggestions?

I can't describe in words how extremely grateful I would be for any help or guidance. I've been beating my head against this for over a month. (Several months actually, since I was previously trying to run SEPP directly through the command line, and that also failed.)

[Stefan (Stefan Janssen)]

Hi Alison,

thanks for reporting this issue. Great that you uploaded the result of the --p-debug output. Reading this output carefully, I found the error KeyError: 'Burkholderia.pseudomultivorans.SUB.INT23.BP2'. My guess is that you either have a mismatch between the entries of your multiple sequence alignment and the reference phylogeny OR maybe the names of your sequences get transformed in inconsistent ways.
Fortunately, you also uploaded your reference artifact. When I check the alignment, I find 122 sequences with headers like Burkholderia.cenocepacia.574 Concatenation of 5 sequences. The newick reference tree also holds a leaf called Burkholderia.cenocepacia.574 but this name misses the Concatenation of 5 sequences part.

Try to get rid of the Concatenation of 5 sequences suffixes in the multiple alignment file and rebuild your artifact. I would assume this should solve your issue.

Good luck!

[aravenscraft (Alison Ravenscraft)]

Omigosh, can't believe I missed that. Thank you SO MUCH Stefan! I fixed this silly error and fragment insertion WORKED. I've finally got a beautiful tree, after months of struggle. I can't thank you enough.