QuPath TMA Dearrayer ignores a column

I have a fluorescence TMA image with 35 cores. For some reason the second column of cores is totally ignored by the TMA Dearrayer, and it only detects 30 cores (screenshot below). I’ve tried playing with the parameters and nothing has worked so far. I have other 35 core images that have worked fine, and I can’t visually see any difference that might cause it to fail.

Any suggestions? Thanks in advance.

Select the column to the left and Add column after in the TMA menu. You then need to select all of the cores and set them to valid (they will be turned off by default). You may also want to check the labeling, but there is also an option in the TMA menu to correct that if you need to.

Thanks for the suggestion. I probably should have mentioned, I’m using QuPath from within a script, so not using the GUI. Hoping to be able to get this working so that it can automatically detect the correct number of cores.

Ah, I don’t think there are currently many options for that aside from adjusting the estimated diameter. @petebankhead may have other ideas though.

The best thing I could think of would be (since you are already scripting), write something to check the X or Y distances between sequential cores, and if a row or column is missing, select and add it in the script. Not something I have tried before though, and would likely take some digging through the code.

Yes, I’m afraid that for TMA studies I would always run the dearrayer in one step, fix up whatever is needed (saving the corrected data file), and then run any subsequent script afterwards using the automatically-dearrayed-plus-fixed data. A fully automated dearrayer that handles more variation would be welcome…