About misjoins correction step

[Giovanni Marturano]

Dear 3D-DNA developers,

thanks for the great tool.

We are using 3D-DNA to scaffold a plant genome of about 500 Mbp and we would like to skip the misjoins correction step and curate the assembly a posteriori with Juicebox.

We used then the parameter “-r 0” and we checked the statistics of the intermediate “assembly.final.fasta” file. We noticed that the number of fragments generated was higher than the number of the contigs in input (we had 911 contigs assembled with PacBio Hi-fi reads that became 1,037)

Do you confirm that the “final.fasta” file correspond to the assembly after the misjoins correction step?

Is it sufficient the “r” parameter to completely skip the corrections  or we should consider other parameters?

Thank you for your help,

 

Giovanni Marturano

[Olga Dudchenko]

Hey Giovanni,

Please see page 17 of the Genome Assembly Cookbook (Figure 4) @ dnazoo.org/methods. (This is a figure from Supp in Dudchenko et al., Science 2017 with small modifications.) The figure should give you an idea of the general 3d-dna workflow. When you set -r 0 this skips misjoin correction, but you still have the step that attempts to do splitting into chromosomes which may result in some contig breakage. If you are for some reason unhappy with it and/or don't want to tweak (as always the advice is to examine the relevant hic/assembly files in Juicebox Assembly Tools) or want to do chrom boundary detection manually you can always just load .0.hic and .0.assembly. Note that you an use the --early-exit option to generate just these files and exit without proceeding with the rest of the pipeline.

Hope this helps,

Olga