Is there anything odd about this residue, e.g. the C–N distance which the PBBuilderuses to find polypeptides.

the distance between CA and N seems to be around 1.5 A for residue 822
`ATOM 6742 N THR B 822 -8.531 -23.264 12.685 1.00 35.76 N

ATOM 6743 CA THR B 822 -9.413 -23.716 13.755 1.00 35.76 C

ATOM 6744 C THR B 822 -10.876 -23.606 13.333 1.00 35.76 C

ATOM 6745 CB THR B 822 -9.186 -22.906 15.045 1.00 35.76 C

ATOM 6746 O THR B 822 -11.761 -24.155 13.993 1.00 35.76 O

ATOM 6747 CG2 THR B 822 -7.945 -23.390 15.787 1.00 35.76 C

ATOM 6748 OG1 THR B 822 -9.020 -21.522 14.711 1.00 35.76 O `

This is AlphaFold2 prediction, so there could be something wrong with the distances actually

Hi @DimaMolod, thanks for reporting this. You can define a distance to the PPBuilder(), which by default is set to 1.8A. For predicted models, this might be a little off. Alternatively, you can use the more robust CaPPBuilder() that checks distances between Ca atoms at a distance of 3.8A, which can be more robust. If you can share the model, it would be great to help us debug!

Hi @JoaoRodrigues and many thanks for your suggestions and help!
Actually, I already uploaded the model as raked_3.pdb.txt in my first comment. Sorry, I changed the extension because pdb files are not allowed. Please let me know if you can download and test this model.
Many thanks again for your help,
Dima

p.s. This happens to AlphaFold2 models regularly, even though fairly seldom. Indeed the distance threshold might need to be slightly adjusted

Ah sorry, I missed the attachment..

p.s. This happens to AlphaFold2 models regularly, even though fairly seldom. Indeed the distance threshold might need to be slightly adjusted

Yeah, roughly 1/3 of the CA-CA bonds in that model are stretched :) Even at 4.0A, you have some violations:

B:GLN3 - B:ASN2 = 4.010 *
B:LEU32 - B:LEU31 = 4.024 *
B:ASP33 - B:LEU32 = 4.090 *
B:GLY90 - B:ARG89 = 4.007 *
B:GLU150 - B:GLU149 = 4.024 *
B:LYS427 - B:GLN426 = 4.022 *
B:MET507 - B:ARG506 = 4.018 *
B:GLU508 - B:MET507 = 4.014 *
B:LEU510 - B:LYS509 = 4.188 *
B:THR513 - B:GLU512 = 4.010 *
B:ILE514 - B:THR513 = 4.271 *
B:TYR589 - B:SER588 = 4.054 *
B:GLY669 - B:THR668 = 4.003 *
B:THR819 - B:MET818 = 4.151 *
B:ARG820 - B:THR819 = 4.073 *
B:THR822 - B:GLU821 = 4.426 *
B:ALA823 - B:THR822 = 4.679 *
B:ASP824 - B:ALA823 = 4.331 *
B:THR825 - B:ASP824 = 4.180 *
B:ASP841 - B:SER840 = 4.117 *
B:CYS862 - B:TYR861 = 4.040 *
B:ARG865 - B:LYS864 = 4.070 *
B:ALA868 - B:PRO867 = 4.132 *
B:GLY877 - B:PRO876 = 4.082 *
C:MET1 - B:LEU894 = 34.938 *  -> This one is obviously a chain break
C:ASP5 - C:GLU4 = 4.012 *
C:GLU6 - C:ASP5 = 4.083 *
C:GLY48 - C:VAL47 = 4.060 *

Whatever you are doing with these models, I'd recommend a very gentle restrained minimization before any analysis.