Similarity based model performance on test set. As control, templates were randomly selected and ranked. Mean and standard deviation of three independent, random runs are shown.

Top-n	Similarity	Random
	Average (%)	Average ± SD (%)
1	17	4 ± 1
3	39	9 ± 2
5	51	13 ± 2
10	71	17 ± 2
20	86	21 ± 2