I am trying to find, is there is any statistical difference between these two sample groups of data.
For example, GCC(A) codon is highly over used in group1 compared to group2. How to get p value for each codon?
Codon(AA) - Group1 - Group2
GCU(A) - 1.976 - 1.886
GCG(A) - 0.645 - 0.570
GCC(A) - 0.664 - 0.610
GCA(A) - 0.715 - 0.934
UGU(C) - 1.006 - 1.107
UGC(C) - 0.994 - 0.893
GAU(D) - 0.809 - 0.812
GAC(D) - 1.191 - 1.188
GAG(E) - 1.163 - 1.029
GAA(E) - 0.837 - 0.971
UUU(F) - 0.619 - 0.573
UUC(F) - 1.381 - 1.427
GGU(G) - 2.053 - 2.235
GGG(G) - 0.473 - 0.470
GGC(G) - 0.843 - 0.898
GGA(G) - 0.630 - 0.397
CAC(H) - 1.385 - 1.248
CAU(H) - 0.615 - 0.752
AUU(I) - 1.441 - 1.333
AUA(I) - 0.339 - 0.134
AUC(I) - 1.220 - 1.532
AAA(K) - 0.647 - 0.647
AAG(K) - 1.353 - 1.353
CUA(L) - 0.691 - 0.385
CUC(L) - 0.712 - 0.977
CUG(L) - 1.472 - 1.747
CUU(L) - 1.124 - 0.891
UUA(L) - 0.985 - 0.912
UUG(L) - 1.015 - 1.088
AUG(M) - 1.000 - 1.000
AAC(N) - 1.422 - 1.476
AAU(N) - 0.578 - 0.524
CCA(P) - 1.000 - 0.916
CCC(P) - 0.190 - 0.135
CCU(P) - 1.876 - 1.831
CCG(P) - 0.933 - 1.118
CAA(Q) - 0.938 - 0.872
CAG(Q) - 1.062 - 1.128
AGA(R) - 1.034 - 1.333
AGG(R) - 0.966 - 0.667
CGA(R) - 0.663 - 0.538
CGC(R) - 1.072 - 0.990
CGG(R) - 0.208 - 0.131
CGU(R) - 2.057 - 2.341
AGC(S) - 0.842 - 0.855
AGU(S) - 1.158 - 1.145
UCA(S) - 0.832 - 0.803
UCC(S) - 1.219 - 1.118
UCG(S) - 0.358 - 0.323
UCU(S) - 1.591 - 1.756
ACC(T) - 1.280 - 1.174
ACA(T) - 0.700 - 0.685
ACG(T) - 0.597 - 0.495
ACU(T) - 1.423 - 1.645
GUU(V) - 1.238 - 1.238
GUG(V) - 0.985 - 1.136
GUC(V) - 0.762 - 0.793
GUA(V) - 1.015 - 0.834
UGG(W) - 1.000 - 1.000
UAC(Y) - 1.234 - 1.091
UAU(Y) - 0.766 - 0.909
UGA(*) - 1.333 - 0.419
UAG(*) - 0.167 - 0.000
UAA(*) - 1.500 - 2.581
I dont know what these numbers are. Provided these numbers are absolute counts, then you can perform a standard chi-squared test.
Looks like these data are real numbers between 0 & 2, roughly. In this is really the case, khi-square test is useless (is even meaningless).
Besides, your questions is not clear : do you want to compare for each line the two values you give --- in such case, you must give more details on how they are obtained --- or to compare the two columns of values --- in such case, the first idea would be something like Student's T test, but to take with caution because your questions is not completely clear...
Minitab, SAS and SPSS manuals will give you the answer. Minitab is, of course, simplest to use aws long as you sight has it.
I think, for what i understood in your question, you have two options. 1.- As many colleagues already pointed out , I do not see why you couldn't run a t-test comparing both groups. This will tell which group have a higher/ lower position (is this what your number means or is the frequency of use of this codon )?.
2.- In the case you were interested in the frequency, or how often each codon is used, you can test the frequencies of each codon by a Chi square, I guess a Pearson Chi-square analysis. your null hypothesis will be that all codon are equally used so the absolute frequency (100%), will be equally distributed in all your different codons. Methods are in Zar, Sokal and Rolf, and almost in every stats book.
I hope this helps!
I think so you are dealing about comparison between two groups values of GCC(A) its not possible because theirs no replication only you have 1 and 1 value of each group.if you make replicate of both groups about GCC(A) then t-test for independence is used to find your desired results........
Thanks guys for all your comments..apologies for the late reply I was away for sometime
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics