Cluster analysis for side chains in amino-acids. Use of a custom measure needed in R^m×3 to apply cluster analysis methodology. Comparison of several distances for 3 - Dimensional structures. Custom matricial similarity measures are proposed. Pioneer approach, quite different from the usual one in the R^p space. Implementation of custom K-means algorithm in Matlab to handle complex 3 dimensional structures.
>> p=genpath('user path\...\Matricial-Distance-for-Cluster-Analysis-of-Amino-Acids');
>> addpath(p)
>> load('cys1010.mat')
>> load('Mc.mat')
>> [ Dnew ] = allin1Cys( cysg,6 );
>> [clustertemp,asigs,ctr ] = kmeans2( Dnew,6,6,0.001 );
>> paintcys(Dnew,6,asigs,Mc );
