Polygon gates with curvy sides after gate xml files being imported into FACSDiva software.

[AaronXtr]

Hi Chloé,
First of all, allow me to briefly express my love for your GateID algorithm. It is a very lucid, practicable and intelligent algorithm.
I am not really sure if this is an issue related to GateID algorithm. We encountered this problem in the last few rounds of gateID-FACS experiments.
After we changed the vertex coordinates to the ones computed by GateID in the xml files exported through FACSDiva and imported the file back through FACSDiva, gates drawn in the software were convex quadrilateral polygons with curvy sides (screenshot of gates attached below). We did switched from log scale to linear scale for vertex coordinates in the xml files, but also changed the boolean parameter to false consistently (screenshot attached below).
We are currently changing the FACS data to log scale from the beginning to see if it could fix the problem. In the meantime, it would be really appreciated if you could offer some instruction and advice.
Again, Love the design of this algorithm!
Best regards,
Aaron Xing
Sample1

Sample2

Sample3

Gate xml file

[chlbaron]

Hi Aaron! I see the problem. I have definitely noticed this in FACS experiments, independently of GateID. I think it's just a funky display behavior. I would make sure to stick to consistent scales and stay away from using a bioexponential scale. Also, are you using the compensated channels or not? Sometimes drawing a gate using the non compensated channel and dragging the same gate in the compensated one can mess things up. Sorry I don't have a sharper answer. But, I'm so glad you are enjoying GateID!

[barveaditya]

Hi @AaronXtr , what Chloe says is correct. GateID should definitely not generate convex gates as I have an explicit routine checking that. One thing you could check is to just plot the GateID generated gate as is, and then check with different transformations that it is undergoing in Diva. That would hopefully help you resolve the issue.