09.09.2014 Views

13th International Conference on Membrane Computing - MTA Sztaki

13th International Conference on Membrane Computing - MTA Sztaki

13th International Conference on Membrane Computing - MTA Sztaki

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

R. Pagliarini, O. Agrigoroaiei, G. Ciobanu, V. Manca<br />

D F ruc6P = {AT P, AMP }, D Gluc6P = {AT P, AMP }, D F ruc16P 2 = {ADP },<br />

D AT P = ∅, D AMP = ∅, D ADP = ∅.<br />

After that, the set R composed by the rules reported in Table 1 has been<br />

obtained by applying the procedure introduced in Secti<strong>on</strong> 4, and the transiti<strong>on</strong><br />

P system Π = (X, R) has been used as starting point to analyze quantitative<br />

causality according with the approach described in Secti<strong>on</strong> 2.<br />

r F ruc6P : F ruc6P → F ruc16P 2 + Gluc6P<br />

r F ruc6P :<br />

F ruc16P 2 + Gluc6P → F ruc6P<br />

r Gluc6P : Gluc6P → F ruc6P<br />

r Gluc6P : F ruc6P → Gluc6P<br />

r F ruc16P 2 : F ruc16P 2 → F ruc6P<br />

r F ruc16P 2 : F ruc6P → F ruc16P 2<br />

r AT P = r AMP : AT P → AMP<br />

r AT P = r AMP : AMP → AT P<br />

r F ruc6P,AT P = r F ruc6P,AMP : F ruc6P<br />

→ AT P + AMP<br />

r Gluc6P,AT P = r Gluc6P,AMP : Gluc6P<br />

→ AT P + AMP<br />

r F ruc16P 2,ADP : F ruc16P 2 → ADP<br />

Table 1. The set of rules modelling correlative causality of the yeast glycolytic network.<br />

Let us c<strong>on</strong>sider v = AT P + AMP . We look for the possible causes for this<br />

multiset, which corresp<strong>on</strong>ds to c<strong>on</strong>sidering two time-series together. To find its<br />

causes, we start by c<strong>on</strong>sidering G = 0 R as a potential cause. Then we proceed by<br />

adding rules to 0 R , <strong>on</strong>e by <strong>on</strong>e, until no more are needed to make v appear. More<br />

details regarding this inductive procedure for finding the cause of a multiset can<br />

be found in [2].<br />

For G = 0 R the c<strong>on</strong>diti<strong>on</strong> lhs(r) ≤ v\rhs(G) does not take place for r =<br />

r AT P ; from the point of view of correlative causality, this corresp<strong>on</strong>ds to saying<br />

that AT P is directly correlated with another time-series therefore it cannot have<br />

an empty cause. We c<strong>on</strong>tinue by adding to the (now discarded) potential cause<br />

0 R rules r which have in the right hand side rhs(r) at least <strong>on</strong>e comm<strong>on</strong> element<br />

with v. The set of these rules is S = {r AT P , r AMP , r F ruc6P,AT P }. For G 1 = G+s,<br />

s ∈ S\{r F ruc6P,AT P } the c<strong>on</strong>diti<strong>on</strong> lhs(r) ≤ v\rhs(G 1 ) remains unfulfilled, since<br />

either AT P or AMP will appear in v\rhs(G 1 ) and both of them are left hand<br />

sides of rules. So we choose G 1 = r F ruc6P,AT P and it verifies that it is a cause<br />

for v. To find the other causes, we look at the discarded causes with <strong>on</strong>e element<br />

(i.e., to the rules from S) and add <strong>on</strong>e element from S to each of them, namely<br />

we c<strong>on</strong>sider all the multisets with two elements of S. By checking all of them<br />

we find that r AT P + r AMP is a cause for v. Note that r F ruc6P,AT P cannot be a<br />

part of a cause with two elements since that rule al<strong>on</strong>e is sufficient in producing<br />

v. Moreover, there is no cause with more than two elements since having three<br />

or more rules in G would mean that <strong>on</strong>e of them does not c<strong>on</strong>tribute to the<br />

appearance of the two elements of v. In the end, we have obtained that the<br />

causes of v are r F ruc6P,AT P and r AT P + r AMP . This corresp<strong>on</strong>ds to the time-<br />

364

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!