convert.sh does not produce a .hic file in the appropriate format for juicebox

[oneillkza]

This is causing various downstream issues, as the next logical step after running SALSA is to manually edit the assembly, and juicebox seems to be the only tool to do that.

See:

aidenlab/Juicebox#958
aidenlab/Juicebox#975

and most likely:
aidenlab/Juicebox#972

[yuzhenpeng]

I met the same problem. How to deal with it.

[oneillkza]

I mean, the answer from the Juicebox team seems to be "Don't use SALSA if you want to edit your assembly in Juicebox".

Or alternately to run the final assembly from SALSA through the entire Juicer workflow to try to get to something compatible with Juicebox. But the Juicer workflow is a nightmare to get running.

[yuzhenpeng]

I mean, the answer from the Juicebox team seems to be "Don't use SALSA if you want to edit your assembly in Juicebox".

Or alternately to run the final assembly from SALSA through the entire Juicer workflow to try to get to something compatible with Juicebox. But the Juicer workflow is a nightmare to get running.

juicebox V 2.X can open the .hic generated by SALSA.

[oneillkza]

Juicebox itself can open the .hic file and visualize it, but the parts of Juicebox for editing the assembly (aka "Juicebox assembly tools" or JBAT) require an "assembly" .hic file, which the file produced by SALSA is not.

Here are some quotes from the Juicebox developer, from the issues I linked in the OP:

Thanks for a detailed description of the issue you are experiencing. Closing since this is not a bug. Looks like you are trying to use JBAT on a sandboxes, non-assembly hic file. Please consult Genome Assembly Cookbook (dnazoo.org/methods) on 3D-DNA tool suite usage. Thanks, -Olga

We do not offer any support for Salsa since we are not developers of that tool: my guess would be that the hic file is not in the right ‘assembly’ format.

[yuzhenpeng]

Okay, Thank you very much.

[c2997108]

I have created a tool to edit a contact map by directly reading SALSA output files (scaffolds_FINAL.agp, alignment_iteration_1.bed and assembly.cleaned.fasta). It only displays at the window size of the first specified resolution (default is 100 Kbp), but it seems to work faster. At least it worked well enough for me.
https://github.com/c2997108/SELDLA-G/blob/master/README.md

[jacquelinemattos]

Has anyone solved this problem?