86
Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery Poster Session I
33-POS
Board 33
Evaluating B-factors for Benchmarking Models of Collective Motion
Edvin Fuglebakk
1
, Nathalie Reuter
1
, Konrad Hinsen
2
.
1
University of Bergen, Bergen, Norway,
2
Centre national de la recherche scientifique, Orléans,
France.
The interpretation of crystallographic B-factors in terms of thermal motion is prevalent in many
branches of protein science, and commonly used to validate or benchmark computational models
of protein motion. This practice implicitly assumes that known limitations to the thermal
interpretation of B-factors can be safely ignored. One common criticism against using B-factors
as a standard for validating modes is that they are influenced by many non-thermal factors.
Another concern is that the thermal component of B-factors arises from motion in a highly
restrictive crystalline environment. This environment is expected to dampen collective motion,
the kind of motion often involved in functional motion like conformational change. I will present
recently published results that reveal potential problems with using B-factors as a model of
thermal motion of solvated proteins (Fuglebakk et al., JCTC, 2013). We have compared the
collective motions of several elastic network models, a kind of protein model commonly
validated and parameterized against B-factors. We obtained collective motion predictions from
elastic network models and compared them with molecular dynamics simulations for seven
solvated protein structures. We find that models that give good reproduction of B-factors are
severely compromised in their ability to recapitulate collective motions. Moreover, we compare
the elastic network models with a null model with restricted collective motion, and find that
models parameterized to reproduce B-factors are in close agreement with this null model. We
therefore find it important to consider the effect of the crystalline environment when interpreting
B-factors, and avoid them altogether when doing quantitative comparisons.
