Alphabet error while annotating phage

[Fabian-Bastiaanssen]

• pharokka version:1.6.1
• Python version:Python 3.10.8
• Operating System:Ubuntu 22.04.2 LTS

Description

I was trying to annotate a circular phage with pharokka, the sequence only contains nucleotides and seems to be fine for other tools. I can't see any reasons why this might throw an error.
sequence_203.txt

What I Did

pharokka.py -f -i all_fastas/fastas/sequence_203.fasta -o all_fastas/sequence_203 -d /data/san/data0/databases/Pharokka/20230827 -t 4 -g prodigal-gv --skip_extra_annotations

2024-02-15 15:57:08.713 | INFO | main:main:363 - Running PyHMMER on PHROGs.
Traceback (most recent call last):
File "/data/san/data1/users/fabian/miniforge/envs/pharokka2/bin/pharokka.py", line 499, in
main()
File "/data/san/data1/users/fabian/miniforge/envs/pharokka2/bin/pharokka.py", line 364, in main
best_results_pyhmmer = run_pyhmmer(
File "/data/san/data1/users/fabian/miniforge/envs/pharokka2/bin/hmm.py", line 34, in run_pyhmmer
with pyhmmer.easel.SequenceFile(
File "pyhmmer/easel.pyx", line 6369, in pyhmmer.easel.SequenceFile.init
File "pyhmmer/easel.pyx", line 6363, in pyhmmer.easel.SequenceFile.init
ValueError: Could not determine alphabet of file: 'all_fastas/sequence_203/prodigal-gv_aas_tmp.fasta's

[gbouras13]

Hi @Fabian-Bastiaanssen ,

I can reproduce this error.

The reason is because the first CDS is

NDDDDDNDDDDNDDDDNDDDDNDND*

which is so repetitive/not-diverse and means pyhmmer can't infer whether it is a protein or nucleotide I am pretty sure and so crashes.

While I look into fixing the error, you can try one of two couple of things:

1. Reroient your circular phage to begin elsewhere, which should hopefully work when you annotate that with pharokka (e.g. with Dnaapler https://github.com/gbouras13/dnaapler - you can do this by passing --dnaapler to pharokka )
2. use --mmseqs2_only to disable pyhmmer

George