Entering edit mode
5.5 years ago
deepti1rao
▴
50
I want to make a consensus genome by replacing variants. I want to replace homozygous variants in the reference. Is it right to replace the heterozygous variants? The ALT allele must correspond to the homozygous chromosome of a diploid organism. That's my understanding. So I am thinking of not replacing such variants.
What is your goal in generating a consensus genome ?
Hi! The goal is to do a reference based assembly. 97% of our reads align to the reference.
So why didn't you use a reference-guided genome assembly (e.g. Velvet, MIRA, Spades, etc..) ?
Currently, I do not have enough data to take up a proper assembly. I have tried velvet de novo. When I mapped the resulting contigs to the reference, I found a lot of misassemblies and the coverage is far less than what my reads cover.
I think all the reference based assemblers initially do a de novo assembly and then map the contigs to the reference. So it must be the same method that I've used. I have only paired end data. No mate pairs. Even the paired end data is pooled from different plants, making it 120 x. I am not sure if this kind of a raw data is good enough for a de novo assembly. Since the reference is so close, I rather want to map reads to it directly and replace the variants in it.
if you want to save information about heterozygosity, you can use IUPAC codes to represent diploid state.
Thanks for this idea! Are two base codes understood and accepted by tools like Bowtie2, samtools etc for future analyses??