I've came across a method in research communication, where different percentages of variation were calculated for a collection of ratio values for some proteins. As seen in the figure, the frequency of (protein) ratios falling within a certain percentage of variation is plotted against the percent variation itself.


Can anyone tell me how this percent variation is calculated?

  • $\begingroup$ We need more information than what is on the graph in your link to help answer the question. $\endgroup$ Commented Sep 17, 2012 at 16:07
  • 2
    $\begingroup$ Hint: the data counts are 24,5,4,1,2,1,1,0,1,1. The cumulative counts therefore are 24,29,33,34,36,37,38,38,39,40. What do you get when you re-express those cumulative counts as percentages of the total (40)? $\endgroup$
    – whuber
    Commented Sep 17, 2012 at 17:42
  • $\begingroup$ @whuber: Do you mean: dat<-c(24,29,33,34,36,37,38,38,39,40); (dat/40)*100 which gives: 60.0 72.5 82.5 85.0 90.0 92.5 95.0 95.0 97.5 100.0 $\endgroup$
    – The August
    Commented Sep 17, 2012 at 22:22
  • $\begingroup$ Yes! More generally, starting with the data in the bars (x<-seq(0.1,1.0,0.1); y<-c(24,5,4,1,2,1,1,0,1,1)), compute plot(100*x, 100*cumsum(y)/sum(y), ylim=c(0,100)) and compare that to your graph. How does it look? $\endgroup$
    – whuber
    Commented Sep 17, 2012 at 22:49
  • $\begingroup$ @whuber Thanks, but I've found in another paper by the same author where percent variation is replaced by percent CV(coeff. of variation) !image I'm still not sure what does percentage of CV mean. The details can be found here $\endgroup$
    – The August
    Commented Sep 19, 2012 at 5:05


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.