variant effects for non-model organisms

[chrishah]

Hi there,

I have assembled a genome de novo. Now I have put quite some effort into gene prediction and lately I have called some SNPs for different populations of the species - was my first attempt to call SNPs so I am new to the SNPcalling field and file formats etc - Please excuse my ignorance.

I am looking for a tool to identify the effects of the variants, whether they lie in predicted exons/introns, whether they change the AA, etc.

I came across snpEFF and was happy at first, but after a closer look I am not sure if it really does what I want. I read also about EnsEMBL Variant effect predictor. As far as I understand these programs rely on annotated genes in databases. As the organism I am working with is a loooonnng way from any of the organisms in the databases I want to use my own predictions (gff3).

Would be great if you could give me any hints or ideas on how to proceed! Thanks!

[Lluc]

I recently succeeded annotating the variants of a non-model species using annovar. It is not straight forward, and I am still checking that I got it right, but I'm satisfied for the moment.

In addition to the list of variants, I prepared 3 files. First, I converted the gff3 annotation to a table like knownGenes from UCSC (name it ending in "_knownGene.txt"). This may be the most difficult step. I wrote a script of my own, that I could share... Second, I generated a dummy file similar to the kgXref from UCSC in which I only filled up the two first columns (known gene ID, and mRNA ID). Name it with the same prefix as before, and ending in "_kgXref.txt". And third, I used the script retrieve_seq_from_fasta.pl, provided with annovar, and the reference genome of my organism in fasta format to generate the fasta file of transcript sequences. Name it with the same prefix and ending in "_knownGeneMrna.fa".

Once you have those three files in the same directory (e.g., /home/database), you run annovar like this:

annotate_variation.pl --geneanno --outfile <outfile> --dbtype knownGene --buildver <prefix_of_your_files> <list_of_variants> /home/database/

Don't forget the final slash in the last argument, indicating the directory where your "database" is.

Good luck.

[chrishah]

Hi Lluc,

THank you very much for your answer and for sharing your approach!!! I appreciate it! I ll have to farmiliarize me a little bit with annovar and will then try your method (might come back to you with another question..).

For now, I have found another approach that might be interesting for you also, if you want to test an alternative to Annovar: snpEFF in fact also enables you to use your own database, which can be created from your draft assembly. The few steps are explained on this website: http://snpeff.sourceforge.net/supportNewGenome.html. After that you just run snpEFF like here: http://snpeff.sourceforge.net/examples.html#ex3. For me it worked right away. I am just trying to assess the results now.

Maybe you wanna try it..

Thanks again for your help! Good luck! much obliged!