Page 62 - Visions of the Future Chemistry and Life Science

World champion chemists 51

ecules’ reactions, or the opponent’s move, may reasonably be predicted for
each possible reaction, but such a calculation will be very difficult. Even if
we assume that this problem is solved, to a sufficient extent for useful
answers to be obtained, then the problem of designing a total synthesis is
still not complete.
Molecules such as bryostatin are synthesised by joining together small
fragments. How many ways can the fragments be joined together? If we
assume that we can buy any molecule with four carbon atoms or fewer,
which is a crude approximation, bryostatin (Figure 3.3) will require about
ten joins, which suggests that there are ten factorial (ten times nine, times
eight, times seven, times six, times five, times four, times three, times two,
which is about three and a half million) strategies to consider. In practice,
the problem is not so straightforward, because many different starting
molecules could be considered, and the adjustments between alcohols and
ketones, and similar transformations, mean that it is necessary to consider
many, many times this number of steps. Two steps for each join might be
a more realistic estimate of the number of steps expected, so the number
of possible approaches is closer to twenty factorial, which is more than a
million million million. Each of these strategies will require the calcula-
tion of the outcome of many reactions, as outlined above, and each of these
calculations is demanding, by the standards of the fastest computers avail-
able today. A complete solution would not be made possible by an increase
in computer power of an order of magnitude, nor even by many orders of
magnitude.
Several orders of magnitude increase in computer power would be
useful to make the calculation of an individual structure rapid, rather than
a major project (for molecules of this size). The conformation searching
problem then requires that many such calculations are performed. To
analyse reactivity many competing reaction processes must be considered
in order to determine the best conditions for a particular transformation.
Many reagents should be considered for each transformation. There are
millions of potential transformations that need to be considered in order
to fully analyse competing strategies for synthesis. To complete these cal-
culations in a reasonable amount of time, which is to say, faster than a syn-
thesis could be accomplished by expert organic chemists without all of this
computational help, will require much faster computers than are currently
available. These calculations will generate an extraordinary quantity of
information which will all need to be analysed. Computers are becoming