Chain and residue numbering issues

[ccgauvin94]

Hi, If I submit a single protein sequence, I get out multiple chains despite the fact that the sequence is just a single chain. Additionally, the chain numbering doesn't make sense. The largest (first) chain, Aa starts with _atom_site.label_seq_id set to 75 which is correct, but the end of the chain is numbered at 311 despite being residue 433. Presumably there are breaks in the chain, but Model Angelo isn't incrementing the seq_id label when it encounters a break.

Additionally, the other smaller chains, despite being from the same sequence, begin their numbering at 1 (when they aren't the first residue), and wind up completely out of register with the actual protein sequence.

Is this the intended behavior? Would it be possible/is it possible to keep the sequence in register with the residue numbering?

[jamaliki]

Hi @ccgauvin94 ,

Is this the result of building a model with model_angelo build or model_angelo build_no_seq. In the case that you are using model_angelo build, is the output file you are looking at output.cif or output_raw.cif. This is intended for the raw output since it also includes chains that are not found in the user sequence, but is not intended behaviour for output.cif, which should just be in register with the sequence provided.

When ModelAngelo encounters a break in the chain and models it as such, it should also increment the label_seq_id for the residues that it is missing.

Could you also please tell me which ModelAngelo version you are running?

Best,
Kiarash.

[ccgauvin94]

Hi @jamaliki - thank you for both the wonderful software, and also taking a look at this issue.

In this case, I ran model_angelo build -v filename.mrc -pf protein_sequence.fasta -o output_directory

In the output directory, I see both the output.cif and outpw_raw.cif files. The raw file has dozens of chains, while the output.cif is the file as I described earlier.

I see this behavior with the latest standalone release (1.0.12), as well as whatever release was distributed with Relion 5.

I could potentially provide you with the map, sequence, and model files, but I'd have to just double check that with the primary investigator.

[builab]

I encountered the same thing when trying to model a ribosome map. Wonder is there any solutions?

[jamaliki]

Hi @ccgauvin94 and @builab ,

I want to apologize for the delay in my response. Could you please provide me with the map and sequence files if you are comfortable with that (email me at kjamali@mrc-lmb.cam.ac.uk)? I can then try to debug the issue.

Best,
Kiarash.

[rabc9]

Hi @jamaliki same issue here with ribosomes, i get fragmented model (well fitted in the map) with changed chain and residue name. Is there any update regarding this issue?

[jamaliki]

Hi @rabc9 ,

Could you provide your sequence and map and give me some info about where you see the issue? I want to see if I can replicate this myself.

[rabc9]

@jamaliki sure ill send you an email with the details. Thank You!

[uddipan-das]

Hi @jamaliki. I am using the ModelAngelo and getting various chains as others are getting above. My model is a tetramer and in principle i should get four chains. The resolution of my map is 2.8A. Could you kindly suggest how should i solve this issue?

[jamaliki]

Could you check to see what version of ModelAngelo you are runing @uddipan-das ? We recently made some changes that should have fixed this in v1.0.13

[csteinmetzger]

Even on v1.0.13 I'm seeing the same issue that others have reported, namely a (ribosome) model that fits well into my map but in which there are dozens of small chains containing hundreds - few thousands of atoms each.