# Is there a way to quickly verify the presence of some SNPs in Fastq files?

Is there a tool that can scan fastq files without assembling them for a custom list of user defined snps?

• Where do fastq come from? How does the list look like? Do you have a reference genome? Please clarify. – aechchiki Aug 19 '17 at 6:35
• They're bacteria sequences from a MiSeq, paired end fastq. They are all the same species and I do have a good reference genome. I'd like to provide a list of genomic coordinates I. The reference genome and find of those snps are present in the new sequences. I'm looking for an easy way just to find the small list I'd like instead of all variants from the reference. – Dan K Aug 19 '17 at 13:45
• could you edit the question and add a sample of your vcf file and the reference genome coordinates file? would be cool to play with, maybe some sed/awk one-liner will do the job – aechchiki Aug 29 '17 at 20:18

Try this pipeline:

1. Clean raw fastq files (trimmomatic)
2. Align clean fastq files to the reference genome (dna-to-dna aligner of your choice)
3. Convert the alignment sam file to bam (samtools)
4. Call SNPs from bam file (samtools)

Here's a start tutorial on this: http://ged.msu.edu/angus/tutorials-2013/snp_tutorial.html

• Thanks, that's a great tutorial. I'm able to call snps and generate vcf files, but I'm looking to quickly query the vcf files for a specific list of snps, preferably listed by reference genome coordinates. I'm not sure on the wording so I haven't been able to find a tool. – Dan K Aug 19 '17 at 13:57

You could use ARIBA, which will map you reads to the regions of your interest (more on that in a second) and then assemble/call variants.

The tool is mostly used to find known variants related to AMR (antimicrobial resistance) from a set of databases, but according to the project's wiki you can build your own database with your region of interest.

Additionally, reference sequences can be either of:

Presence/absence sequences. ARIBA will look for these sequences in the input reads and report any that it finds, and also any variants between these sequences and the input reads.

Variants only sequences. These should have known variant details specified in the metadata file (see below). ARIBA reports only when it finds at least one of the given variants in each of these these sequences. If a sample has one of these sequences, but does not have one of the supplied variants, then it is not reported. If you supply a variants only sequence, but no variant, then the sequence will be removed during sanity checks.

• @DanK, please consider accepting one of the answers if they helped you solve your problem, thanks – mgalardini Aug 22 '17 at 8:22