Biomedical Engineering Reference
In-Depth Information
Table 2.2
Attractors for
melanoma
network
PIRIN
S100P
RET1
MART1
HADHB
STC2
WNT5A
x
1
x
2
x
3
x
4
x
5
x
6
x
7
BAD
0
0
0
0
0
1
1
0
0
1
1
1
1
1
1
0
1
0
0
0
1
GOOD
0
1
0
0
0
0
0
0
1
1
1
0
0
0
1
0
1
1
1
1
0
1
1
0
1
1
0
0
0.1
10mins
30mins
60mins
120mins
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
10
721
744
849
Attractor Cycle Order Number
Fig. 2.2
Average predictor error difference on melanoma attractor data using MiniSat without
modification
only returns one SAT result. We further modified MiniSat to always randomly select
decision variables during the solving process to increase the activity of all variables.
The unaltered MiniSAT uses a heuristic for selecting the next decision variables.
However, this heuristic results in many of the same variables being chosen over
iterative runs of MiniSat. To increase the activity of all variables, we change the
random variable frequency of MiniSat to 100 % (from2%intheunaltered MiniSat
code). This forces MiniSAT to always choose a random variable on every variable-
branch decision. A random variable frequency of
f
% means that MiniSat selects the
next variable randomly
f
% of the time.
To demonstrate the quality of predictor selection using our modified All-SAT,
our algorithm was run on four selected attractor cycle orderings (labeled 10, 721,
744, and 849) using melanoma data from [
3
]. The All-SAT operation was allowed
to run for 12 h (which approximates a complete All-SAT run). In the case of at-
tractor cycle order 721, all cubes were found. In Figs.
2.3
and
2.2
, we compare the
