Tracing genes to the original contigs

[V-JJ]

Hello!

I'm working with highly syntenic genomes of closely related species. I've obtained two genome assemblies.

• ntJoin (default -> no_cut=False) --> (named as Asm_A)
• ntJoin (True) --> (named as Asm_B)

After the scaffolding step we annotated some genes of interest in both assemblies and tried to trace back their coordinates to the original contigs by using the "*.path" file. In principle, should we expect to retrieve the same coordinates for all those genes included in uncut contigs (that is, those contigs that were included entirely in both Asm_A and Asm_B)?

• The thing is that can't retrieve the precise coordinates and consequently can't compare the precise location of my genes of interest between Asm_A and Asm_B.

Here's a schematic representation of my question.

Any advice would be appreciated,
Thanks in advance,

[lcoombe]

Hi @V-JJ,

You should be able to track coordinates between the scaffolded version of the assembly and the original contigs using the path file, yes.
With that, keep in mind that the original contigs may have been reverse complemented before integration into the scaffolds and there may be trimming of input contigs if there were overlaps detected between joined sequences - all of this information is captured in the path file entries.

You could also use the AGP output (if you set agp=True) if that is easier for you to parse. Not that the AGP file will have coordinates starting at 1, while the path file contig coordinates start at 0.

Feel free to give an example or more detail about what you mean about not being able to retrieve the precise coordinates if you are still having issues.

Thank you for your interest in ntJoin!
Lauren

[V-JJ]

Hi @lcoombe !

Indeed, my script didn't consider properly the 0-based coordinates of the *path file. That must be the reason for mismatch of coordinates. I've adjusted the piece of code and now I'm re-running my analysis.

I will tell you as soon as I get the results.
Thanks for answering so quickly!

Vadim