Segmentation fault (core dumped)

[anaome]

Dear Jens,

I have issues completing a full workflow using Taxor 0.2.1.

I have built a small custom index with reference fungal genomes without any problem (about 500 genomes).

I could "search" without issue either (500k ONT reads):
taxor search --index-file myindex.hixf --query-file SRR34636740.fastq --output-file SRR34636740.search.txt --error-rate 0.15

When I run "profile":
taxor profile --search-file SRR34636740.search.txt --cami-report-file SRR34636740.report --seq-abundance-file SRR34636740.abundance --binning-file SRR34636740.binning --sample-id SRR34636740

I get the following output:

Parsing search results...done
Remove matches to nonunique references...done
Remove low confidence references...done
Filter associated references...
Start EM algorithm...started
Initialize prior probabilities ...done
Starting EM iteration 0
Calculate Log Likelihoods ...done
Segmentation fault (core dumped)

To rule out any problem coming from the index, I downloaded the pre-built RefSeq index. Now the "search" step also dies with error:
taxor search --index-file refseq-abfv-k22-s12.hixf --query-file SRR34636740.fastq --output-file refseq.search.txt

checking input ... done!
use syncmer model
Segmentation fault (core dumped)

Last attempt with the Zymo dataset does the same:
taxor search --index-file refseq-abfv-k22-s12.hixf --query-file Zymo-GridION-EVEN-3Peaks-R103-merged.fq.gz --output-file zymo_refseq.search.txt

I tried on different computers, a physical one with 64GB memory and a Google cloud VM with >200GB memory with the same results.

Downgrading (using conda) to version 0.2.0 does not help.

Further downgrading to 0.1.3 works as expected with the pre-built RefSeq index. I had to re-build the custom index with v0.1.3 to succesfuly run "search".

Could it be a version-related issue ?

Thank you !