Morphs: Difference between revisions

A morph is an animation showing the transition between two molecular conformations.

* [http://morphs.proteinexplorer.org Animations in Protein Explorer] developed by [[User:Eric Martz|Eric Martz]] at the University of Massachusetts, Amherst, USA.<ref name='pe'>Protein Explorer: easy yet powerful macromolecular visualization. Martz, E., Trends Biochem Sci. 27:107-9. 2002. [http://www.ncbi.nlm.nih.gov/pubmed/11852249 PubMed 11852249].</ref><ref>[http://morphs.proteinexplorer.org Animations in Protein Explorer] developed by [[User:Eric Martz|Eric Martz]] at the University of Massachusetts, Amherst, USA.</ref>

* [http://www.people.umass.edu/ktheis/ Morphs of T7 RNA polymerase and inhibition of trypsin by serpin] by [[User:Karsten Theis|Karsten Theis]], [[User:Craig T Martin|Craig Martin]], and coworkers, University of Massachusetts, Amherst, USA.<ref name='theisT7'>Topological and conformational analysis of the initiation and elongation complex of t7 RNA polymerase suggests a new twist. K. Theis, P. Gong, and C. T. Martin, Biochemistry 43:12709-15, 2004.  [[http://www.ncbi.nlm.nih.gov/pubmed/15461442 PubMed 15461442]].</ref><ref name='theis'>[http://www.people.umass.edu/ktheis/ Morphs of T7 RNA polymerase and inhibition of trypsin by serpin] by [[User:Karsten Theis|Karsten Theis]], [[User:Craig T Martin|Craig Martin]], and coworkers, University of Massachusetts, Amherst, USA. These are unusual in utilizing [[#Modeling Intermediate Key Frames|hand-modeled intermediate key frames]].</ref> These are unusual in utilizing [[#Modeling Intermediate Key Frames|hand-modeled intermediate key frames]].

[[Image:Mage_hb.gif]]</td></tr><tr><td bgcolor='#e8e8e8'>Toggling between the carbonmonoxy and deoxy conformations of heme in hemoglobin. (This is NOT a morph.) Convergent stereo snapshots from a Kinemage. (Stops after 25 cycles; reload this page to restart the toggling.)</td></tr></table>

Some proteins perform their functions without major conformational changes. On the other hand, some proteins must undergo major changes in secondary, tertiary, or quaternary structure in order to perform their functions. In quite a few cases, investigators have succeeded in obtaining empirically determined structures for a protein in two or more conformations. The challenge for visualization is then to be able to follow the changes in each region between the two conformations.  

About 2004, Karsten Theis, Craig Martin and coworkers released several morphs<ref name='theis' /> that were unusual because they involved [[#Modeling Intermediate Key Frames|hand-modeled intermediate key frames]], and in one case<ref name='theisT7' />, proposed the actual trajectory of the conformational change.

[[Image:Lacrep_anim_small.gif]]</td></tr><tr><td bgcolor='#e8e8e8'>True movie of the morph shown in Jmol elsewhere near the top of this page. (Stops after 100 cycles; reload this page to restart this movie.)</td></tr></table>

All morphs begin with at least two, usually empirical, [[atomic coordinate files]] that represent different conformations of the same macromolecule. In some cases, more than two are available<ref name='vonrhein' />. Simple morphs then involve calculating a series of interpolated intermediate conformations using one of the methods below.

In some cases where very large domain movements occur, or where the authors are proposing the trajectory that actually occurs, it is necessary to model intermediates in which portions of the molecule are positioned by hand. Interpolations are then typically calculated between the initial empirical conformation and the first key frame, between intermediate key frames, and finally between the last key frame and the final empirical conformation. Examples of such morphs are those by Karsten Theis, Craig Martin, and Peng Gong concerning T7 RNA polymerase, and inhibition of trypsin by serpin<ref name='theis' />.

''Chemically possible'' (also sometimes called ''chemically plausible'', ''semi-plausible'', or ''chemically reasonable'') means that, at the least, correct covalent bond lengths and angles are maintained, and that atoms are constrained not to pass through each other. This type of morph requires sophisticated software, but luckily that is freely available and easy to use at the [http://molmovdb.org Yale Morph Server] from Mark Gerstein's group. In addition to the chemically possible nature of the results, this server is very attractive because it totally automates both the creation of the multiple-model morph PDB data file, and also its visualization. The server does the structural alignment of the starting PDB files, and can handle modest differences in sequence between the initial and final PDB files. It then does a linear interpolation followed by some energy minimization to render each frame ''chemically possible''.

The [http://fatcat.burnham.org FATCAT Structural Alignment Server] also produces "chemically possible" morphs. After an optionally flexible alignment (permitting twists at hinge points determined by FATCAT), a linear interpolation is done between the aligned models. Then "the intermediate structures are optimized by energy gradient minimization employing a reduced representation force field."

A minor disadvantage of energy-minimized morphs is that the movement in the animation may be rather jerky. In some cases the molecule resists the transition, so moves slowly at first, then rapidly and "uncomfortably" through the mid-point arriving quickly near the final configuration. In contrast, linear interpolation alone (see next section) produces a smooth animation which some people prefer.

In this method, a series of intermediate models is created in which each atom moves in a straight line from is initial position to its final position. Linear interpolation by itself (without any subsequent energy minimization) is relatively easy to calculate, and often suffices, but also has a number of limitations. When animated, the interpolated intermediate conformations often greatly help in visualizing the differences between the initial and final empirical models. However, bond lengths and angles become unrealistic, domains may artifactually shrink, expand, or distort, and chains may even pass through each other during the interpolated movements. When there are substantial changes in secondary structure or large movements of domains, these artifacts make the eye unable to follow the details of the conformational change, and linear interpolation becomes unsatisfactory.

[[User:Karsten Theis]] has provided a server for linear interpolation between two PDB files: [http://www.bioinformatics.org/pdbtools/morph2 Linear morph between two PDB files]. Use with caution as it has not yet been tested thoroughly.

A very simple program, named '''morph2.exe''', performs linear interpolation and has been freely available since the late 1990's in the ''PDBTools'' package by Eric Martz.<ref>[http://www.umass.edu/microbio/rasmol/pdbtools.htm PDBTools] by Eric Martz include a simple linear interpolation morphing program. This program operates only in MS-DOS (under MS Windows), but the C source code is included. Available within ''Protein Explorer'' are [http://www.umass.edu/microbio/chime/morpher/morphmtd.htm step by step instructions] for making linear interpolation morphs.</ref>. This program, which was used for most of the morphs made by Martz (including the one on this page), requires that the two starting PDB files contain exactly the same atoms in exactly the same order. (Achieving this usually requires some hand editing of the PDB files.) Its advantages include that it is straightforward to include ligand in the morph, or at the end of the morph, and to include multiple protein and nucleic acid chains. [http://www.umass.edu/microbio/chime/morpher/morphmtd.htm Step by step instructions] are available within ''Protein Explorer''. While these instructions were written for Chime, rather than Jmol, everything there applies equally to Jmol except the last section on ''Playback Scripts''.

If you upload your multiple-model morph PDB file to Proteopedia, animating it is as simple as checking that option in Proteopedia's ''Scene Authoring Tool''.

For those who wish to make their own animation scripts in Jmol outside of Proteopedia (e.g. in the [http://bioinformatics.org/jmol-tutorials Jmol Tutorial-Authoring Template]), the script required to animate a multiple-model PDB file in Jmol is much simpler than what was needed in Chime. Here are the commands needed in Jmol: