四种R因子的含义和计算方法

感谢跆拳道会长提供的信息。
http://www.rcsb.org/pdb/staticHelp.do?p=help/advancedsearch/xRayRefinement.html
http://www.phenix-online.org/pipermail/phenixbb/2011-July/017341.html
http://www.phenix-online.org/pipermail/phenixbb/2010-August/015595.html

Refinement R Factors

Searches for structures in the specified resolution range. A histogram icon next to the search box links to a histogram that shows the distribution of resolution in the PDB.

Searches for structures in the specified R factor range. R factor is the measure of the agreement between the structural model and the experimental diffraction data. When solving the structure of a protein, the researcher first builds an atomic model and then calculates a simulated diffraction pattern based on that model. The R factor measures how well the simulated diffraction pattern matches the experimentally-observed diffraction pattern. A totally random set of atoms will give an R factor of about 0.63, whereas a perfect fit would have a value of 0. Typical values are about 0.20. See the "Looking at Structures: Resolution" page (here) for more details.

Note: non-diffraction methods such as NMR do not have a resolution specified.

The R factor search supports four types of metrics:

• R observed - Residual factor R for reflections that satisfy the resolution and observation limits. This quantity includes reflections set aside for cross-validation in a "free" R-factor set. Residual factor R for reflections that satisfy the resolution and observation limits.
• （简单说，就是收数据收到的所以衍射点对应的R因子，包括被reject的点）
• （计算方法：一：

  phenix.refine data.mtz（所有衍射点的mtz） model.pdb  xray_data.r_free_flags.ignore_r_free_flags=true main.number_of_mac=1  strategy=none output.prefix=junk

• 也有人说，在phenix.refine中输出的R VALUE     (WORKING + TEST SET)就是R observed ,

• 不过我表示怀疑，它怎么知道我在数据处理时踢掉哪些点。

• http://www.phenix-online.org/pipermail/phenixbb/2010-August/015595.html
• 
  
• R all - Residual factor R for all reflections that satisfy the limits established by high and low resolution cut-offs. It includes reflections which do not satisfy the sigma cutoff criteria as well as those set aside for cross-validation in a "free" R-factor set.
• （抛掉reject点后，所有点计算的R因子，计算方法同上。见：
• http://www.phenix-online.org/pipermail/phenixbb/2011-July/017340.html
  
• 或者：

  Rall = .05*Rfree + .95*Rwork

  亦或是

  Rall = （1-完整度）*Rfree + （完整度）*Rwork

• 见：
  

• http://www.phenix-online.org/pipermail/phenixbb/2011-July/017341.html
  

  ）
• 
  
• R work - Residual factor R for reflections that satisfy the resolution and observation limits and that were used as the working reflections in the refinement. This quantity does not include reflections set aside for cross-validation in a "free" R-factor set.
• （抛掉reject点，抛掉rfree点后的R因子，计算方法：phenix.refine）
• 在pdb中：R VALUE            (WORKING SET) : 0.2368
  
  
• R free - Residual factor R for reflections that satisfy the resolution and observation limits and that were excluded from refinement to be used for cross validation in a "free" R factor set. R free is a less biased metric, and is typically higher than an R work. If R free is significantly higher than the R work, the structural model may be over fitted.
• （抛掉reject点，然后选择约5%的衍射点计算得到的R因子，计算方法：phenix.refine）
• 在pdb中：FREE R VALUE                     : 0.2630
  

Examples

1. Search for structures with an R free values between 0.18 and 0.24

   "R free": Enter "Between: 0.18 and 0.24

2. Search for structures with an R free value of up to and including 0.22.

   "R free": Leave the left box empty and enter 0.22 into the right box