Jeffrey's X coefficient


A doubt has risen concerning the Jeffrey's X coefficient. According to the formula (#matches / #bands at lane a + #matches / #bands at lane b), the maximum value should be 2 (or 200 when expressed as %). However, in the similarity matrices the maximum number is 100 (the same as the other coefficients). What I would like to have confirmed is that BioNumerics simply normalizes the Jeffrey's X results by dividing them by 2 to facilitate the comparison with other coefficients.


You are right that the Jeffrey's X coefficient turns out to be maximally 2 if not divided by 2. Therefore, the coefficient is known as: Sx = 1/2(Nc/Na+Nc/Nb), with Nc being the number of common bands in patterns a and b. Note that the coefficient is undetermined if a or b has zero bands. Therefore BioNumerics checks for such cases and sets the coefficient to zero if it occurs.


Applicable for: 
Version 1.0 - 7.6


Share this: