This old version of Proteopedia is provided for student assignments while the new version is undergoing repairs. Content and edits done in this old version of Proteopedia after March 1, 2026 will eventually be lost when it is retired in about June of 2026.


Apply for new accounts at the new Proteopedia. Your logins will work in both the old and new versions.


User:Karsten Theis/Rigid morph

From Proteopedia

Jump to: navigation, search

Contents

Problem

Linear morphs give abnormally short bond lengths along the trajectory when the molecule (or a domain) rotates by a lot between starting and ending conformation.

Idea

Morph from one conformation to another by defining domains. For each domain, intermediate conformations are generated by first finding the optimal rotation superposing the two conformations and moving starting and ending conformation somewhere midway (i.e. moving the starting conformation "forward" and the ending conformation "backward"). Finally, do a linear interpolation somewhere midway between the two structures. Instead of atoms morphing on linear paths, they will move on helical paths, preserving bond lengths along the trajectory.

Treatment of translation

If each domain is rotated and translated individually, interdomain connections can become stretched to a ridiculous degree. To prevent that, the morph is done in a hierarchical manner where lower level-domains are tethered via anchors to the higher-level domains. This keeps the distance of connected parts constant throughout the trajectory.

Expanding the functionality

1. Add a debugging flag. If debug = 0, do the entire process. If debug = 1, leave out the linear interpolation. If debug = 2, leave out the translational anchoring and treat each domain independently (center of mass moves on a linear path from start to end conformation).

2. Allow domains with unequal number of atoms (define corresponding atom selection to find rotation operation, and use on all atoms in domain while skipping the linear interpolation for the atoms that are not matched).

3. Decouple rigid body motion from linear interpolation

4. Deal with disparate atoms sets.

Proteopedia Page Contributors and Editors (what is this?)

Karsten Theis

Personal tools