velvet in heterozygous site

[lyz1030]

Hi,

I want to use velvet to do de novo assembly on a particular region (in a bam file) and this region contain a heterozygous site. I intuitively think it should generate two contigs in this heterozygous site, but it only generate one. Can velvet make sense for this problem? If not, which software can work for this?

Thanks

[swbarnes2]

In my experience, velvet will pick one letter, and only print that. You will lose the heterozygosity; velvet is not designed to deal with that.

I'd use velvet to make contigs, then align to your contigs, and call SNPs with GATK or samtools or something. Then you will see the heterozygosity.