Proseg with openst

[yoavnahum21]

Hey!

I'm currently working over an openst spatial data, which is not included over your protocols' list.

I saw there is an option running a custom protocol defining the right arguments running the proseg shell command.

Can you explain how to do this right? it might be cool if you will add another example into the "examples" folder explaining how to run a modified protocol over the proseg correctly.

Thanks in advance, Yoav

[dcjones]

Sure, this is something I should document, but there are a bunch of command line arguments that will sometimes let proseg interpret a format it doesn't officially support if it is some kind of transcript table in csv (or .csv.gz) format.

I think the the minimal set of arguments to make this work are:

• --gene-column: Name of the column holding the gene or feature name.
• --x-column: x-coordinate column name
• --y-column: y-coordinate column name
• --cell-id-column: name of column encoding cell assignment in the prior segmentation
• --cell-id-unassigned: what value in that column represents a transcript being unassigned to a cell.

The other consideration is that proseg assumes coordinates are microns. If that's not the case (e.g. it's in pixels) you also should use --coordinate-scale to give constant to scale the coordinates by to arrive at microns.