blast results load/align structure only aligns first model in NMR ensemble

[Elaine Meng]

For example if you search PDB using 1t2p chain A, then hits 1ija, 2kid, 2mlm are all NMR ensembles. Using doubleclick or the "load structures" context menu item on any of those hits superimposes just the first model. Not serious/urgent in my opinion, can be rescued by simply using Matchmaker "manually." Probably want to match each one rather than using the matchmaker "bring along" option since ensembles are not necessarily matched up with each other to start with.

[Zach Pearson]

This was also the behavior in ChimeraX 1.2.5; can you elaborate on what you mean by using matchmaker manually? If you reply with the sequence of commands you'd use, I can come up with a way to automate them.

[Elaine Meng]

I see that you issue a command like

matchmaker #9.1/A to #2/A

Instead should be something like

matchmaker #9/A to #2/A

...just like when the file has a single model.  Thus it was surprising that you'd actually gone to the effort of specifying one submodel!

[Elaine Meng]

Also most definitely need to include the option "pairing ss" for all of your matchmaker commands, regardless of whether there is an ensemble or not!  Otherwise it won't pay any attention to the chains you specified in the command and might end up matching some other chains if they happen to get a higher score, and will take longer while trying all the pairwise combinations of chains. See
<https://rbvi.ucsf.edu/chimerax/docs/user/commands/matchmaker.html#pairing>

[Elaine Meng]

Although then for the ensemble case it complains about unequal numbers of match and reference chains.  Eric thinks it might be a bug, because matchmaker should be able to figure out when you're specifying a equal number of chains in each match model (to the reference number of chains)  even though there are multiple match models.  Well I'll let you duke it out between yourselves...

If it turns out to be not a bug or not a fixable bug, the alternative is giving multiple commands
matchmaker #9.1/A to #2/A pairing ss
matchmaker #9.2/A to #2/A pairing ss

[Elaine Meng]

Contrary to our expectations maybe it is not required to use "pairing ss" when you have specified only one reference chain and one match chain per model.  Eric says the code is confusing...

[Elaine Meng]

Eric fixed the bug but also determined that it is not necessary to give "pairing ss" if you already gave only one chain per match model and per reference model in the command.  Sorry for the runaround.

[Zach Pearson]

I'm not sure the behavior changed; this is the result of the new matchmaker code.

[Elaine Meng]

To summarize, regardless of what changed or didn't change, it works to just use

matchmaker #2/A to #1/A

instead of

matchmaker #2.1/A to #1/A

...if #2 is the incoming ensemble model (#2.1 ... #2.N).

[Zach Pearson]

Should be fixed by ​this commit.