How to generate the reference genome for heterozygous HiC data?

[xinkwu]

Hi nservant,

The HiC data I used was a hybrid fly, produced by A and B crossed. I used genome A , genome B, and genome combine A and B( I just changed the name of each scaffolds in two genome and used cat* command to combine them together in the same fastq file) as the reference genome, respectively.

The result confused me a lot. I got more mapped pairs, of course, by using combined genome. But I got more multiple pairs alignments and fewer valid pairs by using combined genome.

How to generate the reference genome for my data properly? Do you have any suggestions?

Thanks a lot,
Kai

[nservant]

Hi kai,
If I correctly understood, you generated a kind of diploid reference genome ? with 2N chromosomes ?
In this case, you will indeed expect much more multiple pairs, as all reads that do not overlap a SNPs between A/B will be mapped twice (on the two parental chromosomes).
What do you want to do exactly ? do you want to generate allele specific maps ? to explore the chromosome organization of each parental chromosome ?
If so, you can have a llok to the manual http://nservant.github.io/HiC-Pro/AS.html.
You will need to generate a N-masked genome, where all SNPs between A/B are replaced by a N.
Best
N

[xinkwu]

Yes, I generated a diploid reference genome.
Now the first thing I want to do is to make a quality control of my HiC data, to ensure that my data can support the next step analysis with RNA-seq data.
Because of the hybrid background, the result was not very well when I used A genome as the reference genome only. I supposed there will have more valid pairs when I combine the parental genomes together.
Maybe I could delete the identical sequences between the two genomes(leave one copy only).
Do you think that will be successful?
Best,
Kai

[nservant]

Hi,
I don't know. I never tested HiC-pro with a diploid mapping. I'm sure that this is a good approaches, but I'm only sure that there is many details that HiC-pro will not check, such as repeated reads in common between A/B genome. If you want to perform allele-specific analysis with HiC-pro, please used a N-masked strategy as already mentioned, instead of a diploid mapping.
Best