determining if a SNP is synonymous or not

[nlsullivan]

I work with a pretty uncharacterized bacterium (genome size 4Mb) and I have a collection of SNPs in my mutants. I would like to determine whether or not they are in coding regions and whether they impact the protein sequence.
At the moment I am using Galaxy for my analysis and I have an annotated genome of my bug and a list of SNPs (basepair, parent base call, mutant base call). Is there a way to determine for each SNP whether it is in a coding region and what impact the SNP has on the protein sequence (silent, missense, nonsense) besides looking at each by hand?
I'm pretty new to sequencing and bioinformatics, so appreciate all the help you can give me.
Thank you!

[JohnK]

I like Annovar. They have some helpful conversion scripts and will annotate variations against UCSC or refSeq Gene. You can annotate SNPs or indels. A popular thing to do is download a few gene models and merge them into their union. You can then annotate with the union of the gene model to get a more comprehensive feel for what transcripts your variation might be effecting. Just keep in mind that they're curated differently. Most importantly, Annovar is open-source:

http://www.openbioinformatics.org/annovar/annovar_faq.html

[nlsullivan]

JohnK,
Thanks for your reply. ANNOVAR sounds like it would be great, but it looks to me that it requires that the genome be present on the UCSC genome browser. My genome is not there (and neither are any related or even bacterial genomes). Any suggestions for getting around this?

[JohnK]

You can create your own FASTA with Annovar by going to the homepage, go to gene-based sub-link and then read this:

"I also provide programs to build FASTA sequences for any other genomes for which I do not provide pre-built files.

In summary, in the ANNOVAR package, the FASTA files must be downloaded from the ANNOVAR website directly. The ANNOVAR software handles all these issues for users, via the -downdb argument, so users do not need to how or where to download files. Additionally, I provide a script within the ANNOVAR package called retrieve_seq_from_fasta.pl, such that users can build a FASTA sequence files based on a given genome sequence (such as the 3 billion base pairs in human genome). This is helpful, when users need to deal with a genome for which I do not provide the corresponding FASTA files. "

retrieve_seq_from_fasta.pl

Just make sure you format your fasta file or genome to their specs- remove newlines (or not) if necessary.

[sbberes]

John,
I do this kind of work all the time. To facilitate this process our lab has put together a perl script that takes as input a reference genome sequence in genbank format (which can easily be generated using Artemis or from an automated annotator like RAST) and a list of biallelic SNPs callled relative to the reference genome and ouputs whether the snp is coding or intergenic, and if coding if it is syn of nonsyn, and if nonsyn what the amino acid change is or if it is nonsensense. The perl code requires the bioperl modules. Contact me if you think this script would be of use to you.
Steve Beres
[email protected]

[Giulietta]

If your bacteria is in EnsemblGenomes, you can use the variant effect predictor to calculate the effect of sequence variations on an Ensembl transcript (including any amino acid change). Start with the genomic location and possible alleles. Use the online interface, or the Perl API script:

http://bacteria.ensembl.org/tools.html

Example of input and output:

http://bacteria.ensembl.org/info/website/upload/var.html

At the moment you can use it for these species:

http://bacteria.ensembl.org/info/about/species.html

[Sacrolfur]

You can use Procannot (http://platform.genexplain.com/procannot). It's simple and has user-friendly interface!