Skip to main content
. Author manuscript; available in PMC: 2017 Dec 13.
Published in final edited form as: J Chem Theory Comput. 2016 Nov 10;12(12):6130–6146. doi: 10.1021/acs.jctc.6b00757

Table 1.

Summary of All Protein and Nucleic Acid Systems Investigated with HDBSCAN and Amorim–Hennig Methodsa

biopolymer atoms frames HBDSCAN
Amorim–Hennig
memory (kb) time ([hh]:mm:ss) result memory (kb) time ([hh]:mm:ss) result
MutSα in presence of cisplatinated DNA 1829 5 000 1049724 04:41 1 state per trajectory 16307300 33:28 3 stable states
MutSα in presence of fluoridated DNA 1829 5 000 3080308 27:40 1 state per trajectory 122362552 03:45:46 1 state per trajectory
MutSα in presence of mismatched DNA 1829 5 000 446560 03:32 1 state per trajectory 16755112 55:12 3 stable, 1 transient state
NEMO-CYNZN 28 98 304 1017248 15:46 1 stable state, 75% of frames 7027188 34:22 4 stable states
NEMO-CYS 28 98 304 2380156 05:02:34 42% noise, 2.4% most populated state 64461880 120:00:00 2 frequently switching states
SufC108116 246 1 000 ** 00:33 30% noise, 49% most populated state 2597720 04:37 4 stable, 3 transient states
SufCD108116 671 701 ** 00:12 18% noise, 44% most populated state 146248 02:05 2 stable states
villin headpiece 64 30 605 1570492 34:20 stable segments with periods of instability 33896452 04:25:57 stages of folding
F10 (CaCl2) 197 16 000 204536 04:15 1 dominant state per trajectory 10567036 15:55 1 dominant state per trajectory
F10 (NaCl) 197 16 000 1049908 05:38 74% noise, 1.4% most populated state 8715716 33:48 2 frequently switching states
F10 (MgCl2) 197 1 000 ** 00:10 64% noise, 6.2% most populated state 2608468 09:18 5 frequently switching states
thrombin aptamer 315 25 770 1256036 14:14 27% noise, 17% most populated state 18838956 01:42:49 various levels of compactness
thrombin (KCl) 295 5 000 139136 01:08 29% noise, 14% most populated state 7904572 01:13:29 2 stable, 1 transient state(s)
thrombin (NaCl) 295 50 000 1489776 53:12 1 or 2 dominant states per trajectory 27198820 23:32:15 various loop configurations
a

Atom counts are the number actually used in clustering—heavy atoms for nucleic acids and α carbons for proteins. On three systems (noted with **) HDBSCAN clustering completed too quickly for the distributed computing environment to record the amount of memory consumed. Visualization of structures from additional systems and brief analysis are given in Supporting Information, Figures S11–S20.