|Opened on:||Tuesday, December 30, 2008 - 22:52|
|Last modified:||Monday, July 23, 2012 - 23:07|
Ok, let' face it. rebuild is unquestionably very useful but imo not quite as useful as it might be, in part because of a lack of user control over its behavior.
Suppose for a minute, that a user could apply rules to rebuild aside from the guidance that a rubber band applies to the affected sections.
1. allow players to determine point thresholds for determination of saved output of the rebuild. I would rather walk away *and allow foldit to crunch than to attempt to view the backbone thru the purple shadow to see if the position subsequent to a rebuild iteration is worthy of consideration or just another dead-end position, (which more often than not it is). It is difficult to see the progress as is, and imo, since rebuild already uses some sort of rules to determine "valid" output and has currently no user input aside from "go!" and "Stop!", it would not be contrary to the paradigm. In fact it would further that aim.
2. Allow players to check a box that causes rebuild to discard any "solution" causing hydrophilic residues in the selection to perform translations resulting in inward facing orientation. Again, user-selectable so that those players desiring the transitions could use them those that do not could choose not to allow them...
3. Allow players to check a box that causes rebuild to discard any "solution" causing hydrophobic residues in the selection to perform translations resulting in inward facing orientation. Again, user-selectable so that those players desiring the transitions could use them those that do not could choose not to allow them...
4. Allow the player to select a minimum/maximum allowable negative score transitions as a final constraint to "saved" "solutions" to allow either fine tuning or gross movement to the output
Regarding work involved: I think aside from GUI inclusion of a "Rebuild config" menu tab to the main configuration menu, that the mainstay of this suggestion can be implemented with code already contained within foldit's subroutines and variable list (the original iteration of rebuild used a ?score-better-than-present-score? threshold function to determine whether a new position was valid, ie stored as a file and added to the undo "buffer"), although it would require some additional work adding code associated with the hydrophobic and hydrophilic orientation test. Prescanning the rotamer lookup table and flagging "illegal states" (per philic/phobic rules selected) could also speed up generation of valid, progressive solutions.
Note that #3 and #4 could be used whether for an aqueous or lipid-based protein so for future use it would likely still have value...
thanks for listening .... again