Feb 12

Question 1

Why don’t we get 100% of completeness of crystal structure?

Answer 1

Becuase points at edge of sphere don’t diffract.

Question 2

What is multiplicity of crystal structure data?

Answer 2

It is same as redundancy, how many times on average was reflection measured per point.

Question 3

What is Rmerge value?

Answer 3

It is each measurement compared to the average intensity, so it is how similar values are to noise essentially. Once you get above anywhere from 60 to 80%, then you need to truncate the data set. This value will weight it based on # of measurements, so if only have 1 reflection of a point then less impact

Question 4

What is cc1/2 value?

Answer 4

It is correlation coefficient. It is when divide observations randomly into 2 groups, then compare each measurement to average. If below any values give lower than 0.3, we don’t include it as it is too much noise.

Question 5

For the signal to noise (intensity/std deviation), what is the min value you can go too before it is too much noise?

Answer 5

You can go down to 2, if have higher value could include more data as res is already high.

Question 6

If you had high signal to noise ratio, and high cc1/2 value, and high redundancy, what is a reason you might not shrink the high resolution shell of your measurements (measurements with the highest resolution0

Answer 6

If the completeness is low, so for example entire data is 96 percent, but high res shell is 67 percent, if take out more then it is even lower