Minor internal improvements have been made.
Shape TK 3.6.0¶
OEFlexiOverlapFunc and OEFlexiOverlay have been updated to allow users to use a Shape Query as the reference system. This allows OEFlexiOverlapFunc SetupRef and OEFlexiOverlay SetupRef to accept a Shape Query.
A new method, SetupFlex, has been added. This method can be used to set up overlap calculation functions when molecules are expected to be flexible and their self overlaps are not fixed to the initial value.
A new method, SortedOverlay, has been added as an alternative to Overlay. This method provides a desired number of top hit results instead of all of the hits and is especially suitable when working with molecules with excessive numbers of conformers.
Two additional control flags, GetRigidOverlay and SetRigidOverlay, have been added to the OEFlexiOverlayOptions class. These flags indicate if rigid overlay optimization between the fit molecule conformers and the reference system should be performed prior to the flexible optimization.
Major bug fixes¶
An issue in OEFlexiOverlay that caused a TanimotoCombo value greater than 2.0 when a molecule was flexibly fitted against itself has been fixed.
Minor bug fixes¶
The default value for the
colorfunction in OEFlexiOverlapOptions during OEFlexiOverlay has been changed from
exact. This ensures that color overlaps are reported when flexible optimization is performed with default parameters.
An issue has been fixed with OEFlexiOverlay that was causing shape Tanimoto to be occasionally reported as larger than 1.0 after a flexible optimization.
The legal range of values for the color score multiplier has been added to the OEFlexiOverlapOptions and OEOverlayOptions option classes.
OEColorFFParameter now allows users to select explicitmillsdeannorings properly as the color force field.
The acceptable ranges for the number of optimization iteration steps have been defined for OEFlexiOverlayOptions and OEOverlayOptions.
Newer implementation of Hermite code had limits on lambda parameters of the expansion. In rare cases, these were affecting the lambda parameter search (Newton optimization stepping outside the limits). Removing the limits fixed the previous problems with finding lambda parameters.