Calculates a mutual information score (MI) based on the probability of joint occurrence of events and corrects it with the average product which is subtracted from the MI score.
mip(align, fileHelix= NULL, diag= 0, fileCSV= NULL, gap_val= 0.8, z_score= TRUE)An object of class 'align' created by the import.msf or the import.fasta function from a sequence alignment
A string of characters that indicates the file containing the positions of the anchor residues in the sequence alignment. To be used for the analysis of GPCR sequences. Default is NULL.
A numeric value indicating the score of the diagonal elements in the scoring matrix. Default is 0.
A string of characters indicating the name of the csv file where the output matrix will be saved. Default is NULL.
Numeric value indicating the gap ratio at a given position for this position to be taken into account. This value must be between 0 and 0.8. Default is 0.8, which means that positions with more than 80 percent of gaps will not be taken into account.
A logical value to perform a Z-score normalisation of the covariation matrix (TRUE) or not (FALSE). Default is TRUE.
A list of two elements : a matrix containing the correlation scores for each pair of positions and, optionally, a second matrix with the Z-scores
The MIp score at position [i,j] has been computed with the following formula :
$${MIp(i,j)} = MI(i,j) - \frac{MI(i,\bar{j})MI(\bar{i},j)}{<MI>}$$
with :
\({MI(i,j) = \sum_{x,y}^{ } p_{x,y}(i,j) ln\frac{p_{x,y}(i,j)}{p_{x}(i)p_{y}(j)}}\)
\(MI(i,\bar{j}) = \frac{1}{n-1} \sum_{j \neq i}^{ } MI(i,j)\)
\(MI(\bar{i},j) = \frac{1}{n-1} \sum_{i \neq j}^{ } MI(i,j)\)
\(<MI> = \frac{2}{n(n-1)} \sum_{i,j}^{ }MI(i,j)\)
and where \(p_{x,y}(i,j)\) is the frequency of the amino acid pair (x,y) at positions i and j.
N.B. this formula has been widely applied in the field of sequence covariation but favors pairs with high entropy.
Dunn SD, Wahl LM, Gloor GB. Mutual information without the influence of phylogeny or entropy dramatically improves residue contact prediction. Bioinfor;atics 2008;24:333-340. Martin LC, Gloor GB, Dunn SD, Wahl LM. Using infor;ation theory to search for co-evolving residues in proteins. Bioinformatics 2005;21:4116-4124.
# NOT RUN {
align <- import.fasta(system.file("msa/toy2_align.fa", package = "Bios2cor"))
#Creating MIP object
mip <- mip(align)
# }
Run the code above in your browser using DataLab