On 3 March 2017 at 19:51, Mahdi, Sam <sam.mahdi.846@xxxxxxxxxxx> wrote:
Hello everyone,
This isn't as much a problem with relax as it is a general question
regarding S^2 values itself. When comparing different values of S^2 between
proteins, what difference is considered significant. Due to the lower
values I know say an average S^2 value of .1 is a big difference in the
flexibility of one protein to another, but what about say an average
difference of .01, is that still considered a significant change in
considering the relative dynamics of one protein to another? (I.e. if say,
protein A has an average S^2 value of 9, and protein B has an average S^2
value of 8.9, could you make the statement Protein A has significant
dynamic differences, more rigid, when compared to Protein B. Or could you
make the statement due to only a difference of .01, there isn't much of a
overall dynamic difference between the 2 proteins)?
Hi Sam,
Sorry for not getting back to you earlier, I have been extremely busy
lately! This is really a statistical problem. And it depends on the
quality of your parameter error estimates. If you have two values and
two associated errors, then you can look at hypothesis testing / ANOVA
statistics to tell you if the two values are statistically different
(i.e. 1st year Uni stats). Note that your error estimates should
include the variability of the diffusion tensor, as that will account
for probably around 80% of the influence on the relaxation data, and
without this variability a S2 parameter comparison is meaningless.
There is no rule for S2 differences so you need to rely on basic
statistical tests.
Regards,
Edward
Re: S^2 differences between proteins