ambiguous reactions
17 views
Skip to first unread message
James Lynch
Mar 21, 2024, 4:24:34 PMMar 21
to kappa-users
Section 3.7 of the Kappa Manual is a discussion of ambiguous reactions. It concludes with
"Molecular ambiguity can be very costly. ... Costs like these are unavoidable, but some clever computer science could improve efficiency." Does anyone know of a Kappa add-on or extension that implements this clever computer science?
"Molecular ambiguity can be very costly. ... Costs like these are unavoidable, but some clever computer science could improve efficiency." Does anyone know of a Kappa add-on or extension that implements this clever computer science?
Jean Krivine
Mar 22, 2024, 11:04:49 AMMar 22
to James Lynch, kappa-users
Hi James,
Indeed molecular ambiguity is not easy to deal with. As far as I know, nothing that improves significantly on what Kappa does has been implemented. The best you can do is replace an ambiguous rule A(...),B(...) -> A(...),B(...) by refinements that remove the ambiguity (a la BNGL).
In some cases this is doable w/o too much combinatorial explosion.
Suppose that A(x),B(y) should bind at rate k2 if they are not already bound and k1 if they are in the same complex, and assume they are n ways (n paths) that can connect A and B via other sites than x and y. The problem is enumerating LHS that prove that A and B are *not* already bound is quadratic in n. Whereas the LHS that prove that A and B are already bound amounts to enumerating "only" n rules.
One thing you can do is to approximate the expected behavior by writing n rules for the unary case:
r11 = A(...,x),..P1...,B(...,y) -> ... @ k1
...
Indeed molecular ambiguity is not easy to deal with. As far as I know, nothing that improves significantly on what Kappa does has been implemented. The best you can do is replace an ambiguous rule A(...),B(...) -> A(...),B(...) by refinements that remove the ambiguity (a la BNGL).
In some cases this is doable w/o too much combinatorial explosion.
Suppose that A(x),B(y) should bind at rate k2 if they are not already bound and k1 if they are in the same complex, and assume they are n ways (n paths) that can connect A and B via other sites than x and y. The problem is enumerating LHS that prove that A and B are *not* already bound is quadratic in n. Whereas the LHS that prove that A and B are already bound amounts to enumerating "only" n rules.
One thing you can do is to approximate the expected behavior by writing n rules for the unary case:
r11 = A(...,x),..P1...,B(...,y) -> ... @ k1
...
r1n = A(...,x),..Pn...,B(...,y) -> ... @ k1
where Pi are explicit paths that connect A and B
and only one rule :
r2 = A(x), B(y) -> ... @k2 for the pseudo binary case.
The rate of r2 will be overestimated by all instances of r2 that are indeed unimolecular. But this will be negligible for high numbers of A and Bs.
and only one rule :
r2 = A(x), B(y) -> ... @k2 for the pseudo binary case.
The rate of r2 will be overestimated by all instances of r2 that are indeed unimolecular. But this will be negligible for high numbers of A and Bs.
On Thu, Mar 21, 2024 at 9:24 PM James Lynch <jly...@clarkson.edu> wrote:
Section 3.7 of the Kappa Manual is a discussion of ambiguous reactions. It concludes with
"Molecular ambiguity can be very costly. ... Costs like these are unavoidable, but some clever computer science could improve efficiency." Does anyone know of a Kappa add-on or extension that implements this clever computer science?
--
You received this message because you are subscribed to the Google Groups "kappa-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kappa-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/kappa-users/48418789-2424-40cd-8032-bd69721b856cn%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages