clustering time grows linearly with .dist file size, right?

Theory behind mothur
1

Hi,

Please remind me whether this is correct: Clustering time grows quadratically with the number of sequences, but linearly with the size of the distance file. This is because the size of the distance matrix already grows quadratically with sequence number, right?

I am talking “cluster” command here.

Thanks,
F

2

I think that should be correct for cluster.classic. Although, we do cluster differently since we turn everything into a sparse matrix. I think it might actually be a bit faster and definitely uses less memory.

3

Thank you!