[cc of message to jlmccarthy at lbl.gov]
John, some time back you asked for some feedback on the MultiMap. Yesterday
I finally got an imaginative and demanding user (Steve Tanksley) to sit down
and go through it with me. Our reactions below are based on his first-time
interaction, with me as operator, after my having used MultiMap several times
and read the docs but not actually run Jean's g2pmap.def example. My table
definition multimap-dem.def is given at the end of this message.
First, he/we were very happy with some aspects of the current version.
("current": Actually this is still xace 1.8.) Specifically,
- the tie-lines between the loci on the two maps, and between them and the
- the connection back to the Tablemaker window so that clicking a locus
on the MultiMap highlights it in the table
- the pickability of the locus name in the table, so that double-clicking
it pulls up one of the gMaps it exists on.
Problems, suggestions, fantasies:
- The default scales, with each map being independently scaled to full size,
are a good choice. But Steve also felt it would be good to have a
way to easily show both maps on the same scale. Maybe a button labeled
"Same scale" or "Proportional"?
- We were able to zoom the two maps independently with the "Zoom Active"
setting, but we never did figure out how to select which map was "active".
- We wanted to be able to display all loci that were on either map, with only
those common to both maps connected by tie-lines. So far the best I've
been able to get is this: By changing the query condition on Colonne 1 from
Mandatory to Optional, I get all the loci that are on the other map
listed in the table and shown on the multimap; but the position of the
non-common loci is shown as 0.0 in Colonne 2 and that's where they're all
plotted on the multimap. A mess.
- The main suggestions were about how to friendlify the query mechanism.
These were guided by Steve's conviction that in crop plants we will never
need to compare anything but gMaps. From that perspective it seems
unnecessarily complicated to have the Tablemaker table involved at all, if
one could double-click on a locus name in the multimap to jump directly to
= Aside: Steve wondered why the multimap can't be just two gMap windows
fused, with tie-lines added. This would have the advantage that the full
power of the main query mechanism could be used to build a keyset of loci
to be shown/hidden/highlighted with the Show Selected Objects etc.
buttons. In any case, it would be good to have two handy buttons for
Show Common Loci, and Show All Loci.
The basic query is just to pick two Linkage_Groups ("Chromosomes" in
some people's models) to be compared. Very simple question, should be
executable very simply.
= I suppose the most easily implemented approach would be to have the user
filter out a keyset containing the Linkage_Groups to be compared using the
Template, or maybe a query, then choose "Compare maps" from the menu.
= Easier to use would be for a command (e.g. "Compare maps") to call up a
"Map Chooser" window listing all the available Linkage_Groups and allowing
two or more of them to be selected.
= Steve's vision was a window that would list chooseable Linkage_Groups
in an array by homeologous sets. e.g:
Tomato Potato PepperMap1 PepperMap2
1 1 1 1
2 2 5 5
... (or whatever the homeologues actually are)
Since of course we don't yet know what the homeologues are in all species,
this would have to be an array generated by loadable data. For example
there might be a new class ?Homeologue_set:
Map_Chromosome ?Map ?Linkage_Group
Homeologue_set : "2"
Map_Chromosome Tomato 2
Map_Chromosome Potato 2
Map_Chromosome PepperMap1 5
Map_Chromosome PepperMap2 5
I hope this is helpful, or at least comprehensible.
// Spread sheet definition for the ACeDB software
// User: matthews
// Date: Thu_22_Apr_1993, Time: 15_18_54