Issues with SnpEff Assuming Circular Chromosomes in Eukaryotic Genome Analysis

Question

[ekirsch]

Hello BioStars Community,

I have been encountering an issue with SnpEff where it assumes circular chromosomes during the annotation of a eukaryotic genome (Savannah Sparrow), which should have linear chromosomes. The genome I'm working with is a lift-over from a closely related species, and I'm unsure if this contributes to the problem.

When running the snpEff.jar dump command to inspect the database contents, I receive output suggesting that SnpEff is creating "mirror genes" due to the assumption of circular chromosomes. Here is an example of the output:

Gene 'GENE_exon-XM_058815638.1-4' spans across coordinate zero: Assuming circular chromosome, creating mirror gene at the end.
Gene        : Gene_JAKOOL010000202.1:-60081-58974
New gene    : Gene_JAKOOL010000202.1:186076-305131
Chrsomosome : Chromosome_JAKOOL010000202.1:1-246157

...

Total: 4 added as circular mirrored genes (appended '_circ' to IDs)

Here's an example of the dump command used:

java -jar /spack/apps/linux-centos7-x86_64/gcc-8.3.0/snpeff-2017-11-24-ouvgtdabu7an6qtjqs47fzrrbx5jxq6b/bin/snpEff.jar dump -v bPasSan1 -c snpEff.config

I'm looking for insights on the following:

1. Is there a specific setting required to correct this assumption for eukaryotic genomes in the config file?
2. Any suggestions on how to edit the GFF file or modify the SnpEff configuration to prevent this incorrect circular chromosome assumption?
3. If it is only adding 4 circular genes, will this have a big effect on downstream analyses?

Thank you for any guidance or advice you can provide!