I agree completely for error messages regarding errors that the user can fix.
However that is not the case with this error, which was an “unexpected” error — perhaps I will add a “contact developer” message to those errors to make this clear. Thanks, Nick On 25 May 2018, at 16:22, "herman.schreu...@sanofi.com<mailto:herman.schreu...@sanofi.com>" <herman.schreu...@sanofi.com<mailto:herman.schreu...@sanofi.com>> wrote: Dear Nick, Thank you for your reply. I am looking forward to test a new Pandda version that is hopefully compatible with the latest CCP4 version. Concerning the error messages: I think they are very useful for developers, to trace back in the code what had happened, but not for users, who need information how to change the input files and/or parameters to fix the problem. However, Charles Ballard send me an email that the problem could also have been caused by old .pyc files and gave instructions how to remove these. If the alignment is to get a superposition, you could consider aligning on sequence numbers, which is trivial and which should be identical for all data sets. Best regards and have a nice weekend! Herman Von: Pearce, N.M. (Nick) [mailto:n.m.pea...@uu.nl] Gesendet: Freitag, 25. Mai 2018 15:59 An: Schreuder, Herman /DE Cc: CCP4BB@JISCMAIL.AC.UK<mailto:CCP4BB@JISCMAIL.AC.UK> Betreff: [EXTERNAL] Re: [ccp4bb] Pandda problems Hi, Yes, that’s not a very useful error message — sorry. I am still in the process of intermittently rewriting the pandda code to be …better. I don’t think you’ve necessarily done anything wrong — pandda has not been extensively tested on running on series of structures that are not identical (we normally refine one reference structure against all datasets). However, your timing is pretty good, because I will (today) release a new version of pandda that should work with the latest version of ccp4. I think it should fix the particular problem that you have. I’m just running tests now, but on Monday you should be able to update your ccp4 to the newest update and then update your version of pandda using pip, following the instructions on the website (https://pandda.bitbucket.io<https://urldefense.proofpoint.com/v2/url?u=https-3A__pandda.bitbucket.io&d=DwMGaQ&c=Dbf9zoswcQ-CRvvI7VX5j3HvibIuT3ZiarcKl5qtMPo&r=HK-CY_tL8CLLA93vdywyu3qI70R4H8oHzZyRHMQu1AQ&m=pIVoj_vY0iZI7b1jCo9sKCstLJBflXDxeehLO9PEB9U&s=aiR_cyVICaVVLw8M5Uh4UIcY4TROi_XWnKii9xF65KA&e=>). If that doesn’t work on Monday, let me know if you have any more problems. Thanks, Nick On 25 May 2018, at 14:36, herman.schreu...@sanofi.com<mailto:herman.schreu...@sanofi.com> wrote: Dear bulletin board, I am trying to run Pandda on a set of about 50 data sets. I asked our system manager to roll back to CCP4 update 047 and by using unique and refmac, I could fix the problem of a few missing low resolution reflections. However, then Pandda crashes apparently during alignment of the structures with the following – for me not very helpful – error messages: Failed to align dataset x10 Traceback (most recent call last): File "/DD/Applications/sb/ccp4/ccp4-7.0/lib/python2.7/site-packages/pandda/analyse/functions.py", line 68, in run alignment = model.align_to(other_hierarchy=other.hierarchy, method=method, require_hierarchies_identical=False) File "/DD/Applications/sb/ccp4/ccp4-7.0/lib/python2.7/site-packages/giant/dataset.py", line 96, in align_to self.alignment = align_structures_flexible(mov_hierarchy=self.hierarchy, ref_hierarchy=other_hierarchy, **kwargs) File "/DD/Applications/sb/ccp4/ccp4-7.0/lib/python2.7/site-packages/giant/structure/align.py", line 254, in align_structures_flexible l_ali = align_chains_flexible(chn_mov=chn_mov, chn_ref=chn_ref, altlocs=altlocs, cutoff_radius=cutoff_radius) File "/DD/Applications/sb/ccp4/ccp4-7.0/lib/python2.7/site-packages/giant/structure/align.py", line 285, in align_chains_flexible chn_ref_cr, chn_mov_cr = common_residues(chn_ref_cb, chn_mov_cb) File "/DD/Applications/sb/ccp4/ccp4-7.0/lib/python2.7/site-packages/giant/structure/select.py", line 79, in common_residues assert (max(m_seq_1)<len(alignment.a)) and (max(m_seq_2)<len(alignment.b)), 'Something has gone wrong: selecting residue index greater than chain length' AssertionError: Something has gone wrong: selecting residue index greater than chain length The same message is repeated for 47 more data sets. Does anyone has an idea what the problem is? Did the rollback to update 047 fail? All pdb files have the same sequence, but there may sometimes be one or two residues missing at the N- or C-terminus. Also in some cases, some side-chains may be missing. There is only a single protein chain and a few calcium ions. In most cases, the calcium ions have the same chainID as the protein (as in the reference pdb file), but sometimes the calcium ions have a separate chainID. Any help will be greatly appreciated! Herman
