Skip to main content

Table 1 AUC and elapsed time for 200 queries in the ASTRAL 95% data set.

From: Fast and accurate protein substructure searching with simulated annealing and GPUs

      

95% confidence interval

Method

Platform

Restarts

Elapsed time

AUC

Standard error

lower

upper

SHEBA

CPU

-

25 h 22 m

0.931

0.004

0.924

0.938

SA Tableau Search

CPU

4096

142 h 42 m

0.930

0.004

0.923

0.937

SA Tableau Search

GTX 285

4096

4 h 11 m

0.930

0.004

0.923

0.937

SA Tableau Search

Tesla C1060

4096

5 h 40 m

0.930

0.004

0.923

0.937

SA Tableau Search

CPU

128

4 h 18 m

0.921

0.004

0.913

0.929

SA Tableau Search

GTX 285

128

0 h 08 m

0.920

0.004

0.912

0.927

SA Tableau Search

Tesla C1060

128

0 h 11 m

0.920

0.004

0.912

0.927

QP Tableau Search

CPU

-

157 h 51 m

0.914

0.004

0.906

0.922

LOCK2

CPU

-

208 h 09 m

0.912

0.004

0.904

0.920

TableauSearch

CPU

-

1 h 11 m

0.858

0.005

0.848

0.867

TOPS

CPU

-

1 h 11 m

0.855

0.005

0.845

0.864

VAST

CPU

-

14 h 26 m

0.855

0.005

0.846

0.865

IR Tableau

CPU

-

0 h 01 m

0.854

0.005

0.844

0.864

YAKUSA

CPU

-

4 h 14 m

0.827

0.005

0.816

0.837

  1. Area under the ROC curve (AUC) and elapsed time for different methods on the 200 query set against the ASTRAL SCOP 95% sequence identity non-redundant database. The table is sorted by AUC descending.
  2. In the Platform column, either a GPU card is described, or CPU, which is a single core of an AMD Quad Core Opteron (2.3 GHz, 32 GB RAM) running Linux.
  3. SARF2 is not present in this table, as it is unable to process a data set this large, and DaliLite is not present as it exceeds the CPU time limit.