How to track residues/chains deleted?

[pcbve1]

MOSAICS is designed to delete specific residues and atoms within the program and calculate scores from there. This has been successfuly implemented. however, actually tracking the order in which chains are deleted is challenging.

Chain order:

Currently, MOSAICS uses a chain_order defined by mmdf indexing of chains, which often differs from the order in ChimeraX, where templates can be manipulated. There is no good way to track the order of chains, especially if the chain order is randomized using the randomize_chain_order parameter. by dumping the chain_ids parameter to a json file using MosaicsManager.model_dump(), this may be solved, as this parameter is no longer a private variable like _chain_order was.

Residue order:

MOSAICS uses a sliding window iterator to delete certain residues along a chain, and uses the mmdf residue_ids column to index the residues. However, to index the residues, one needs to recreate the sliding window iterator themselves. In the case of skipped residues, this can be quite confusing. A way around this is to store a variable in ResidueTemplateIterator, a residue_ids variable that can be used to define the start and stop of a residue deletion window.

[mgiammar]

There is no good way to track the order of chains, especially if the chain order is randomized using the randomize_chain_order parameter.

The AlternateTemplateResult class should contain all necessary information about which chains/residues/atoms were removed for that specific alternate template. The class MosaicsResult has a list of alternate template results held under the parameter alternate_template_results: list[AlternateTemplateResult].

The attributes chain_ids & residue_ids of an AlternateTemplateResult instance describe exactly the (chain, residue) pairs which were removed. These IDs/indexes match with the PDB file which was fed into MOSAICS (or what is contained within the DataFrame, if modified separately from the PDB).

Figuring out chain ordering for a MosaicsResult should then iterate through the alternate results and look at which chains were removed.

MOSAICS uses a sliding window iterator to delete certain residues along a chain... In the case of skipped residues, this can be quite confusing.

I'm not aware of cases where specific residues are skipped by the window function and/or iterator (besides the end, which may have a lesser length than the desired window size). Do you have an example for such?

Also, see above comment on the AlternateTemplateResult class which will always describe which chains/residues were removed for that alternate template. Atom IDs are also indexed similarly, but are not tied to a specific chain/residue and are rather the global IDs for the PDB structure under investigation.