Buz,
> Does anyone know the exact meaning of the output from the
> first rigimol processing step?
Those tables were used for diagnostic output during development of RigiMOL.
> For instance, if I give the rigimol program two structures
> that are very similar (structure #1 and structure #2), the
> program identifies several domains and gives the following output:
>
> Rotations and Translations
> Domain RMSD Axis Angle Translation Distance
> 1 0.19 ( 0.17, 0.26, 0.91) -0.00 ( 0.00,-0.00,-0.00) 0.00
> 2 0.17 (-0.65, 0.13, 0.74)0.35 ( 0.04, 0.06,-0.03) 0.08
> 3 0.21 ( 0.33, 0.33, 0.82) -0.19 (-0.03,-0.03,-0.04) 0.06
> 4 0.20 ( 0.49, 0.41, 0.61)0.27 (-0.02, 0.02,-0.06) 0.06
> 5 0.14 ( 0.40, 0.04, 0.92) 16.41 (-0.74, 0.36, 0.20) 0.85
> 6 0.20 (-0.51,-0.47, 0.48) 23.92 (-0.39,-0.09,-1.21) 1.27
> 7 0.51 ( 0.61,-0.15, 0.82) -67.57 ( 1.27, 0.40,-0.49) 1.42
> 8 0.04 ( 0.11,-0.14, 0.99) 113.72 ( 0.58,-0.07,-0.09) 0.59
> 9 0.11 ( 0.02, 0.36, 0.85) -65.37 (-0.28, 0.80,-0.62) 1.05
>10 0.05 (-0.67, 0.55, 0.36) 48.20 (-0.44,-1.14,-0.26) 1.25
>11 0.41 ( 0.73, 0.13, 0.85) -88.51 (-0.22, 0.13,-0.11) 0.28
>12 0.02 (-0.17, 0.88,-0.39) -131.04 ( 0.07,-0.42,-0.43) 0.60
>13 0.02 (-0.46, 0.96,-0.05) -86.21 ( 0.07, 0.03,-0.92) 0.92
>14 0.00 (-0.74, 0.23, 0.76) -111.08 (-1.16, 1.00,-0.61) 1.65
>
> 1. How meaningful is this output, if you'll excuse the rather
> vague question?
They're meaningful in the same sense as would be a set of straight-line
segments fit to a smooth continously-varying curve. There are an infinite
number of sets of lines that could result from such a fit, and the same is
true of RigiMOL's so-called "domains".
In other words, the precise domain boundaries are for the most part
arbitrary, but every domain set will, together in the aggregate, approximate
the underlying transformation (the only "meaningful" part) to a greater or
lesser extent based on the fitting and clustering parameters.
Because RigiMOL's (sole?) intended purpose is to assist with PyMOL-based
visualization of conformational differences between structures, little to no
effort has been invested into figuring out which domains sets are rigorously
better or worse. For purposes of morphing and visualization, any resonable
domain set seems to work.
> 2. On a related thought, what exactly do the columns in the
> output mean? I can imagine that the RMSD is the root mean
> square deviation between domain 1 in structure #1 and domain
> 1 in structure #2.
It's the internal RMSD after fitting domain 1 in structure #1 to domain 1 in
structure #2. A higher value means more motion occurs within the domain.
> 3. What is the Axis column, and how is the axis defined?
> Also, what about the angle, and how is it defined? Is it the
> rotation of the domain about that axis from structure #1 to #2?
It's the axis direction and angle about which the domain must be rotated
during superposition.
> 4. I'm guessing that the translation column gives the
> translation vector of the center (maybe center of mass) of
> the domain from structure #1 to #2. Am I right, and is the
> center the center of mass?
It is the translation component (only) of the superposition -- not mass
weighted.
Cheers,
Warren
