Blastprotein: Visual fixes

[Zach Pearson]

Paraphrased from Elaine's email:

1) Make buttons more like other dialog: i.e. get rid of "Blast" button and instead have Help, Apply (run calculation and keep dialog up), Reset (reset to initial default parameters), Close (dismiss), OK (run calculation and dismiss). See for example Hbonds (in menu under Tools... Structure Analysis).

(1a) Title bar should say "Blast Protein" instead of "Blastprotein"

(1b) Default matrix is wrong, should be BLOSUM62 not BLOSUM45

(2) title bar should say "Blast Protein Results" instead of "BlastProtein Results" (i.e. just insert a space).

(2b) Seems like the default name for the search is None, so one can end up with multiple results panels all with the same name. Is that a problem? (e.g. for sessions?) Maybe the default name should be automatically based on the query, e.g. set to the uniprot ID or structure-chain spec or first N chars of raw sequence, or sequential (bp1, bp1, bp3 ...)

(3) Initial sort order should be highest to lowest score. I was surprised to see it was not in this order. Do you need Score to be shown in the table to sort on it? If so, add it to the suggested default columns in #6 below, next to Evalue.

(4) There is some "processing hits 0/N" line at the bottom that seems to be taking space for no reason. I've never seen it say anything other than 0/N. What processing? maybe it's non-0 at some point but should be removed when it's no longer showing any progress.

(5a) More column name tweaks.
Replace all "Num" and "Nr" with "#" (or at least "Num Polymers" -> "Nr polymers" so they're consistent)
Chain Species -> Species

(5b)
See also #6, column width. Where the widest thing in the column is the header, maybe we can make it shorter, e.g.
Resolution -> Resln

See also below, if we get rid of some of the column choices, that allows further simplification.

(6) Default columns. We definitely don't want ALL of the checkboxes turned on by default. Also is it possible to control the left-to-right-order of columns? I recommend the following:
=-=-=-=-=-=
For AlphaFold, defaults (and left-to-right order):

Name
Evalue
[Score, if we need it for sorting]
Description

Often some of the column values (Title, Species, ...) are blank but can be discerned from the Description value. The contents of Chain Sequence ID are inconsistent and perhaps that column should be nuked (for AlphaFold searches only, I mean). See ticket #5216. Sometimes the same hit is listed multiple times, #5217.

=-=-=-=-=-=
For PDB/NR, defaults in the order:

Name
Evalue
[Score, if we need it for sorting]
Description
Resolution
Ligand Symbols

Several hits are missing Structure ID even though it is obvious from Name what they should be. But I favor nuking the Structure ID column anyway.

For PDB searches, Chain Sequence ID is a UniProt identifier so I suggest renaming that column UniProt. For AlphaFold searches, above and in #5216 I suggested nuking Chain Sequence ID... if that doesn't conflict with the PDB suggestion for this column.

In Chimera, we offered fewer columns than those I see the current ChimeraX dialog. I lean toward paring down the ChimeraX choices similarly, which would allow simplifying some headers. E.g.

• Date – structure deposition date (thus omitting Revision Date and Publish Date)
• Authors – structure authors (thus omitting Citation Authors)
• PubMed – PubMed identifier of literature reference, if any
• UniProt – UniProt identifier, if any, for hit chain (this is currently called Chain Sequence ID)

Others I consider nukable are Replaces, Status, and Structure ID (as mentioned above, is obvious from Name).

(7) Column widths. Initial width should be based on widest item, but needs to be capped to some maximum (although the user should be able to adjust manually to wider later) and the text should wrap instead of showing an ellipsis. Typically there very long values in several different columns (Title, Description, Species, Ligand Name, etc.) so the simplest would be a single max for all, or if it's not a nightmare, perhaps a somewhat larger cap for Title and Description than for the others. Eric can probably help with that.

(8) I'm thinking maybe the results list should be above the columns for all the checkboxes. The checkboxes section could be shown/hidden with a Columns button like in Chimera (Zach probably you should try the Chimera one). Also would be nice if somebody could save their checkbox configuration as a preference, another thing that Eric can probably help with.

[Zach Pearson]

(It's a checkbox)
​1
​1a
​1b
​2
​2b (as part of #5267)
​3
​4
​5a
5b Prohibitively complex, I think.
6: ​1670e2 and ​2e6a1e
​7
​8

Revisions:
​Sort by smallest to highest evalue
Possibly include column index