comparison of three protein samples compositions

[ramanj]

I am having SPSS, Matlab and R ( all three on windows).
I have proteins sets from three different organisms performing the same type of function. I want to compare them in terms of their amino acid composition and see if there are some amino acids over-represented in the three sets with respect to each other. I want to get graphical visualisation as well as numerical information of the differences.

Does someone have ideas on this ?

For two samples, I have seen Welch two sample t-test using R.

[kingkong]

You may try to run ANOVA on each amino acid for three organisms by using SPSS, i.e. if you have 6 amino acids, then there will be 6 ANOVA tests performed. If the ANOVA result for a particular amino acid is not significant, then you can assume the amino acids are the same among the three sets. If the ANOVA result for the amino acid is significant, then you can further study the post-hoc multiple comparisons (i.e. Set 1 vs Set 2, Set 1 vs Set 3 & Set 2 vs Set 3) to see which set or sets are significantly different from others.

For visual presentation, you may try the error bar plot.

See if it works!

[ramanj]

suppose if the case is even more complex and I want to compare the dipeptide compositions of the three sets. Then do I have to run 400 ANOVA tests ?

Can there be any simpler test which would automtatically do it for all the variables and pick out the ones which are most significantly different ?

[kingkong]

Then you may use One-way MANOVA, it will run the comparisons of all 400 variables for you. If it is an exploratory study, you may also consider Discriminant Analysis, which will show the distinct significant dipeptides relating to each set.

[ramanj]

Let me talk of the simple case of amino acid composition. I think that rather than ANOVA, Kruskal-wallis test would be better, because I do not expect the amino acid compositions to be normally distributed.

Please suggest.

[BioSULTAN]

thanks

[Deborahdhw]

ÿþ[

ÿþ[

ÿþ[

[Deborahdhw]

ÿþ[

ÿþ[

ÿþ[

[Deborahdhw]

ÿþ[

ÿþ[

ÿþ[

[Deborahdhw]

ÿþ[

ÿþ[

ÿþ[