The number of amino acids in a protein is important to predict their size. I used coding DNA sequence (CDs) to predict the average of the protein size of E.coli k12 from a sample provided by the seqinr R package. The size of each sequence is stored in variable x. I used two histograms to fit the distribution of this variable and their logarithm.

enter image description here enter image description here

I suggest that the log linear distribution is more reasonable.The average E can be calculated then:


  • Which average do I have to use?
  • Is there a correlation between the base composition and the predicted average?
  • 1
    $\begingroup$ You seem to have two questions but I don't really understand them. Here is my try to rephrase your questions. 1) What standard theoretical distribution best fit the observed distribution? This is a question of statistics, not biology. 2) Is there a correlation between GC skew and protein length? This question has little to do with the rest of the post. $\endgroup$ – Remi.b May 10 '16 at 20:07
  • 1
    $\begingroup$ Which average do I have to use? They are both the same, since you're using the exact same data. It's only the presentation that's different, as you're comparing a linear plot to a log plot. $\endgroup$ – MattDMo May 10 '16 at 21:12
  • 1
    $\begingroup$ @MattDMo Not necessarily. OP might want average log weight which is not the same as log of average weight. $\endgroup$ – Memming May 10 '16 at 21:40
  • 2
    $\begingroup$ Can you elaborate your question? I am having a hard time understanding what you are exactly asking for. $\endgroup$ – Memming May 10 '16 at 21:41