BWA for virus

[mathew]

I have a short read data from Hiseq (originally doen for mouse but has integrtaion of virus). I do have FASTq file of refrence genome. I was wondering what parameters I need to change if I use BWA/ Bowtie to align it. Can I use these or is there any other option

Thanks

[swbarnes2]

You probably have a fasta of the viral reference, not a fastq.

But bwa will handle that fine.

For the best accuracy, what you want is a reference fasta with the whole mouse genome and the viral genome, and you align everything to that. But aligning to a mammalian genome is slow. Aligning to just your virus will be a whole lot faster, but you might get some mouse reads being forced to align to your virus. Maybe set it to only allow a single mismatch, then see if there are SNPs, fix your reference to match, and go again until you fix all those discrepancies.

[nanto]

It depends how much coverage you have with those fastq reads. BWA will create substrings either way. Considering reads are of good quality you might want to use suggested parameters by the program. Worth of changing is definitely -z to value 100

How good reads are mapped you probably know, that % of mappeing from flagstat in samtools is not a good statistic for this. You should check NM values, number of hits per read, number of uniquely mapped reads and total number of mapped bases to choose finally best parameters for mapping. Try increasing penalties in b,q,r