Electron cryomicroscopy: Difference between revisions
| Line 13: | Line 13: | ||
[http://tinyurl.com/method-vs-resolution tinyurl.com/method-vs-resolution].</ref>. For comparison, the median resolution of X-ray crystallographic entries in the PDB has been 2.0 Å for many years<ref name="mvr" />. When resolution improves by a factor of 2, the available data (to support the coordinate model) goes up by a factor of 8. For example, a 2.4 Å resolution structure is a great improvement over a 3.0 Å resolution structure because the number of available measurements doubles. | [http://tinyurl.com/method-vs-resolution tinyurl.com/method-vs-resolution].</ref>. For comparison, the median resolution of X-ray crystallographic entries in the PDB has been 2.0 Å for many years<ref name="mvr" />. When resolution improves by a factor of 2, the available data (to support the coordinate model) goes up by a factor of 8. For example, a 2.4 Å resolution structure is a great improvement over a 3.0 Å resolution structure because the number of available measurements doubles. | ||
A direct comparison between the quality of cryo-EM structures and crystal structures is not possible. In crystal structures, the electron density is calculated from measured structure factors and from calculated phases (in the most extreme case, half of the information is not available from experiment). Even in cases were there are experimental phases (e.g. from multiple isomorphous replacement), these are typically not available to the full resolution. Especially at resolutions below 4 Å, X-ray structures are prone to model bias in the absence of experimental phases. Cryo-EM does not have this limitation, so low resolution structures can still carry reliable information, for instance about conformational changes of known structures. For example, a 4 Å crystal structure solved by molecular replacement (i.e. no experimental phases) is not as reliable as a 4 Å cryo-EM structure. If atomic models of the components of a structure are available, like in the structure of the insulin receptor depicted at right, it is possible to build an atomic model from 4.3 Å resolution data (with some | A direct comparison between the quality of cryo-EM structures and crystal structures is not possible. In crystal structures, the electron density is calculated from measured structure factors and from calculated phases (in the most extreme case, half of the information is not available from experiment). Even in cases were there are experimental phases (e.g. from multiple isomorphous replacement), these are typically not available to the full resolution. Especially at resolutions below 4 Å, X-ray structures are prone to model bias in the absence of experimental phases. Cryo-EM does not have this limitation, so low resolution structures can still carry reliable information, for instance about conformational changes of known structures. For example, a 4 Å crystal structure solved by molecular replacement (i.e. no experimental phases) is not as reliable as a 4 Å cryo-EM structure. If atomic models of the components of a structure are available, like in the structure of the insulin receptor depicted at right, it is possible to build an atomic model from 4.3 Å resolution data (with some patience, the EM density will appear upon clicking <jmol><jmollink><text>here</text><script>isosurface s_one color blue "http://proteopedia.org/wiki/images/2/2b/Map6SOF.jvxl" mesh</script></jmollink></jmol>). | ||
</StructureSection> | </StructureSection> | ||