Calculates the amount of similarity (or dissimilarity based on how you are looking at the data) within a dataset.
Value
three matrices are returned.
Tanimoto.Indices: pairwise Tanimoto coefficients for all samples
Overlap.Counts: number of overlapping samples for each area of interest. Must be read column-wise because the number of counts is for each row area within the column.
Overlap.Ratios: ratio (aka fraction) of number of overlapping samples for each area of interest. Must be read column-wise because the ratio is for each row area within the column.
Details
The number and ratio (aka fraction) of overlapping samples within each area of interest along with the Tanimoto coefficient/index are calculated. The returned Tanimoto coefficients/indice matrices are symmetrical around the diagonal, but the count and ratio matrices are not. They can be interpreted two different ways.
Count and Ratio Matrices
The diagonal of the count and ratio matrices contain the number of samples containing the area of interest.
Within each column, the returned counts or ratio values indicate the number (ratio) of samples within each area that contains both areas (the column and the row). The returned column values are always to be compared to the column of interest. The returned count will never be more than the diagonal value of the column.
Each row, especially for the ratio matrices, should be interpreted as the similarity of each column to each other. For example, the ratios within a row
Within each row, the returned counts or ratio values indicate the number (ratio) of samples within each area that contains both areas (the column and the row).
Author
Emilio Xavier Esposito emilio.esposito@gmail.com (https://github.com/emilioxavier)