[hoffmann]

hi all -- first post 

say i can predict chemical reactions in vivo.  i generate a list of chemical reactions:

7+160   <-->   27+48
26+134   <-->   25+555
42+2291   <-->   97+1118
231+417   <-->   5379+11437
555+1137   <-->   170+315
16+160   <-->   48+1352
26+134   <-->   217+555
48+4479   <-->   160+12624
197+1081   <-->   1353+5125
5+26   <-->   6+25
37+84   <-->   188+1242
188+2051   <-->   1242+5519
5+26   <-->   6+217
37+84   <-->   188+2051
188+1242   <-->   2051+5519
22+197   <-->   118+1146

the numbers are just arbitrary compound IDs.  of this list of predicted reactions, the first 3 actually happen in vivo.  the rest of the reactions haven't been observed (yet?).  what does it mean to say the following:

how many reactions does it take for a falsely predicted reaction to become a true, observed chemical reaction when compared to a database of known chemical reactions?

can this be viewed as a sequence of reactions leading to a real reaction?  any insights?

[Borek]

I would say two.

Let's say you have predicted
A + B = E + F

and database contains
A + B = C + D
and
C + D = E + F

Obviously starting with A, B you can get E, F.

But the question is ambiguous to me, there are so many hidden assumptions floating around that I wouldn't trust the answer.

[hoffmann]

i suppose this assumes that every species pair can react with every species pair.  obviously not every type of reaction can happen...

[mascott]

It is also always possible that E+F interact/interfere with A+B, inhibiting the reaction from continuing. This does require some A+B to produce E+F in the first place...

Remember that in a reaction pathway, an energetically unfavourable reaction can be combined with an energetically favourable one to drive the reaction forwards.

[hoffmann]

so let's say we know the all the true reactions that exist.  for the reactions in the list that we haven't observed (false reactions), what could we find out about them if we did the following:

find the smallest number of true reactions that contain all the compounds in a single false reaction.

this seems like it would give the same information as the previous method but without the assumptions.