2
$\begingroup$

I want to cluster protein conformations by dihedrals angles. My point is an n-dimensional vector, where is n - number of dihedral angles. I think I can't use Euclidean distance for distance metric because of the distance between +179 degrees and -179 degrees is 2 degree, not 358.

Can you suggest some clustering algorithms and distance metrics for such tasks? Maybe some algorithms from geo points clusterization.

$\endgroup$
1

1 Answer 1

2
$\begingroup$

You have circular variables. One way to deal with them is to create two variables for each: sin(alpha_1), cos(alpha_1), sin(alpha_2), cos(alpha_2), ... Then you can use regular clustering algorithms.

There exists custom algorithms for circular data. Examples: Unsupervised clustering of multivariate circular data

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.