Illegal instruction (core dumped)

[Edouard94]

Hello dear @nextgenusfs and @hyphaltip,

First thank you for the very useful tool you created and for keeping improving it.

I was excited to try funannotate2 and compare the results to the ones obtained with fun1, but I got the error 'Illegal instruction (core dumped)' when calling the tool (after creating a new conda env as recommended).

Do you think my system won't be able to run it, here are the cpu specs:

Thank you for your help regarding this.

Best wishes,
Edouard

[nextgenusfs]

Shouldn't be an issue with your system I wouldn't think but maybe one of the conda packages. What commands did you run and try?

[Edouard94]

Hello @nextgenusfs,

I followed the cmd recommended for installation. I get the error message when I try to see the help message.

For example funannotate2 --help.

I also tried to run the cmd funannotate2 train \ -f ~/GregarineStuff/Lgryllorum_Genome_analysis/Funannotate/Lgryllorum_sort.fasta \ -s "Leidyana gryllorum" \ -o train -cpus 10 \ -t ~/GregarineStuff/Lgryllorum_Genome_analysis/Funannotate/species_train/training/funannotate_train.pasa.gff3

But I continuously get the core dumped error - I tried to remove and recreate the conda env it did not change the issue. Fun1 is running correctly in its own env.

I also tried to install funannotate2 in my fun1 env (and added miniprot and table2asn), but it did not fix the issue.

[nextgenusfs]

Hmm strange as core dump not normally an error for python module. You could install f2 in the f1 environment with pip.

[Edouard94]

Yes I tried that already, I will try to see if I can get around my issue and will let you know if I find a solution. Thanks!

[nextgenusfs]

Hi @Edouard94, I've made quite a few changes to this tool and its dependencies. If you wouldn't mind trying again that will help to know if this is still an issue. Trying to stamp out some issues prior to pushing the tool to pypi and bioconda. I believe this below should be sufficient.

mamba create -n funannotate2 gfftk gapmm2 minimap2 miniprot snap "augustus==3.5.0" glimmerhmm diamond trnascan-se buscolite
conda activate funannotate2
python -m pip install git+https://github.com/nextgenusfs/funannotate2.git

[edwhisnant]

Hey Jon @nextgenusfs ,

I was able to install F2 using your instructions above.

The installation seems to have worked when I run:

funannotate2 -h
usage: funannotate2 [-h] [--version] {install,clean,train,predict,annotate,species} ...

Funannotate: eukaryotic genome annotation pipeline

Commands:
  install     Install funannotate2 database files.
  clean       Find and remove duplicated contigs, sort by size, rename headers.
  train       Train ab intio gene prediction algorithms.
  predict     Predict primary gene models in a eukaryotic genome.
  annotate    Add functional annotation to gene models.
  species     View and manage species in F2 database.

Help:
  -h, --help  Show this help message and exit
  --version   show program's version number and exit

[nextgenusfs]

Here are some instructions for a full test if you have the bandwidth!

First setup the databases, you can add the ENV variable to your conda environment like this:

conda activate funannotate2
conda env config vars set FUNANNOTATE2_DB=/path/to/funannotate2-db
conda deactivate

Then we can install the databases.

conda activate funannotate2
funannotate2 install -d all

After that you could test with Aspergillus nidulans ref genome, download the reference genome from NCBI:

wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/011/425/GCF_000011425.1_ASM1142v1/GCF_000011425.1_ASM1142v1_genomic.fna.gz

Then we can just run f2 train on that data like this:

funannotate2 train -f GCF_000011425.1_ASM1142v1_genomic.fna.gz --cpus N  -s "Aspergillus nidulans" --strain FGSCA4 -o anid_f2

After training is complete you can run f2 predict on those training data.

funannotate2 predict -i anid_f2/ --cpus N

Finally you can add functional annotation:

funannotate2 annotate -i anid_f2/ --cpus N

[nextgenusfs]

Oh, and of course genemark is not installed with the above method due to licensing, but if it is in your $PATH than it will run it alongside.

[edwhisnant]

Fantastic. I will run this test.

Does the genome need to be run through RepeatMasker (or another masking program) prior to funannotate train?

[nextgenusfs]

Not specifically required, this test is soft masked. If the genome is not soft masked, f2 will run pytantan to do just a quick job.