3D-DNA parameters

[Michael Ma]

Hello!

I'm using 3D-DNA to assemble my genome to chromosome level, and I use this pipeline:

~/run-asm-pipeline.sh -r 2 \

--editor-repeat-coverage 10 \

--editor-coarse-resolution 5000 \

~/xxx.fasta \

~/merged_nodups.txt

but the output xxx.1.hic seems not good, with so many blocks not in the diagonal line:

however, if I adjust "normalization" option to "coverage", It seems better:

What's more, there is a white band on the right corner of the heat map, which has low coverage less than 10, but why those regions were not removed for using "--editor-repeat-coverage 10" ?

Thanks in advance! 

[Amrita Raj]

Dear Michael,

I am a PhD student currently working on HiC data in plants. I am new to Hi-C data analysis and don't have much background in bioinformatics. I came across your question and would like to know how do you use Juicer for HiC data mapping, normalization, compartment calling etc.. . Could you please share some Juicer scripts/github links that will guide me through stepwise? I would really appreciate your guidance!

Thanks in advance!

[Olga Dudchenko]

Hello Michael,

As you can see from the coverage track your data has huge coverage biases. This may be due to the data or due to the assembly (if it contains a substantial number of alt haplotypes, for example). Normalization view attempts to correct for those biases  so that you can get a better view of your data. White band is stuff with no aligned Hi-C data. Without signal it cannot be anchored and was left unanchored by the algorithm. 3D-DNA puts everything unanchored after the main assembly, i.e. to the lower right corner.

Olga

[Michael Ma]

Hi Amrit，

I'm new to HiC analysis too!  I just follow the manual in https://github.com/aidenlab/juicer/wiki  and the Genome Assembly Cookbook  of Aiden Lab , you can find it here: https://www.dnazoo.org/methods . Hoping this helps !

[Amrita Raj]

Hi Michael,

Thank you for the github link and the cookbook!

I see in your graph you have eigenvector values (A/B compartments) for your chromosomes. Could you please tell me what coverage did you use to look at these structures for your genome of interest?

Best,

Amrita