Nicolas Fusseder wrote: > I have an application which models protein sequences based on template > structures which are found by blasting the query sequence against all > sequences from pdbaa. The entire modeling process works fine, only when > one of the templates structures found via blast was solved with Cryo-EM > the modelling process fails with the following error (sorry for the nasty > formatting): ... >> P1;1D3E1 > structureE:1D3E:1 :1:285 :1:MOL_ID 1; MOLECULE INTERCELLULAR > ADHESION MOLECULE-1; CHAIN I; FRAGMENT:MOL_ID 1; FRAGMENT 1 - 185; > ORGANISM_SCIENTIFIC HOMO SAPIENS; ORGANISM_CO:28.00:-1.00 > APVAAYVDEVLNEVLVVPNINQSHPTTSNAAPVLDAAETGHTNKIQPEDTIETRYVQSSQTLDEMSVESFLGRSG > CIHESVLDIVDNYNDQSFTKWNINLQEMAQIRRKFEMFTYARFDSEITMVPSVAAKDGHIGHIVMQYMYVPPGAP > IPTTRDDYAWQSGTNASVFWQHGQPFPRFSLPFLSIASAYYMFYDGYDGDTYKSRYGTVVTNDMGTLCSRIVTSE > QLHKVKVVTRIYHKAKHTKAWCPRPPRAVQYSHTHTTNYKLSSEVHNDVAIRPRTNLTTV*
Modeller 8 only recognizes the following alignment types as denoting structures: structure structureX structureN structureM Anything else (e.g. structureE) is considered a raw sequence, and no PDB is read in. This has already been corrected for the next Modeller release; as a workaround until then, you can just edit your alignment file to replace structureE with structureX or similar.
Ben Webb, Modeller Caretaker