You've got an antibody which is outside the norm of antibodies - according to the automated annotation in ROSIE, the framework region in your light chain is too short (28 verus a standard 32 to 34). Due to the way ROSIE Antibody is set up, it can't apply the automated proceedure to thread this onto the standard backbones.
Past that, given that your antibody is outside the norm for antibodies, you're likely going to have to run the antibody modeling proceedure manually. So you'd download Rosetta locally, and then follow the RosettaAntibody proceedures outlined in https://doi.org/10.1038/nprot.2016.180 You're likely going to need to tweak the alignments and settings to be able to get Rosetta to work with your "short" antibody, or to correct the misalignment that the automatic tools are doing.
The issue is with the light chain, not the heavy chain. Check all the sequences (all the loops, all the framework regions, for both chains) that are reported in the output of your ROSIE job, and see if they match up with what you think they should be.
Comparing different system for deliniating the start and end points of the loop (the various system in the different papers I mentioned) is a good idea, as any conflict there may point to what the issue may be with the way Rosetta is modeling things.
You've got an antibody which is outside the norm of antibodies - according to the automated annotation in ROSIE, the framework region in your light chain is too short (28 verus a standard 32 to 34). Due to the way ROSIE Antibody is set up, it can't apply the automated proceedure to thread this onto the standard backbones.
The first thing to check is if this antibody light chain does indeed match what it's supposed to. I'd highly recommend reading https://doi.org/10.1016/j.jmb.2010.10.030 and https://doi.org/10.1006/jmbi.1997.1442 and looking through the website https://www.bioc.uzh.ch/plueckthun/antibody/ to check your antibody sequence against what is "normal" for antibodies, to see if there's something special/funky going on with your particular light chain.
Past that, given that your antibody is outside the norm for antibodies, you're likely going to have to run the antibody modeling proceedure manually. So you'd download Rosetta locally, and then follow the RosettaAntibody proceedures outlined in https://doi.org/10.1038/nprot.2016.180 You're likely going to need to tweak the alignments and settings to be able to get Rosetta to work with your "short" antibody, or to correct the misalignment that the automatic tools are doing.
I am counting 32 aa between CDR2 and 3 unless I am using a different system to assign the CDRs. Is this probably what the issue is?
The issue is with the light chain, not the heavy chain. Check all the sequences (all the loops, all the framework regions, for both chains) that are reported in the output of your ROSIE job, and see if they match up with what you think they should be.
Comparing different system for deliniating the start and end points of the loop (the various system in the different papers I mentioned) is a good idea, as any conflict there may point to what the issue may be with the way Rosetta is modeling things.