Predictive Workshop Minutes


University of Minnesota Biodegradation Prediction Workshop June 20, 1998

Biodegradation focuses on specialized catabolic reactions for the metabolism of environmental pollutants.
A biodegradation prediction project could have multiple goals:
  1. Is a compound likely to reach a given end state, such as "80% metabolized to CO2 within 14 days by activated sludge".
  2. Prediction of one possible chemical route - a sequence of reactions known to be possible from organic chemistry.
  3. Prediction of one possible biochemical route - a sequence of plausible enzymatic reactions. Plausibility is: either the reaction is similar to one catalyzed by a known enzyme or an unknown enzyme may be proposed based on the chemical reactivity of the substrate.
  4. One possible route to intermediary metabolism.
  5. Prediction of all possible chemical and biochemical routes.
  6. Prediction of all possible chemical and biochemical routes to intermediary metabolism.
  7. Prediction of routes using first principles at some level of theory - e.g. qualitative organic chemistry molecular mechanics quantum mechanics molecular dynamics
  8. Evaluate novel biochemical reactions for chemical feasibility at some level of theory.

Pathway prediction may run into combinatorial problems, since each reaction can initiate large numbers of potential pathways. However, you might do combinatorial things once and store them for repeated use.
The system is not just the computation, it is the expertise incorporated into the system and the expertise needed to evaluate the output.
Two questions on expert predictions:
  1. How well do expert predictions agree with each other (precision)?
  2. How well do expert predictions agree with experiment (accuracy)?

Examples of Expert Input Big Dumb Rules (Heuristics)
  1. If it is big, break it into pieces at the most vulnerable bond.
  2. Amidases and esterases are in all bacteria and there are often several of them present, and thus they are the first reactions to consider.
  3. If it repeats, chew it repeatedly from one end.
  4. When it is small enough, take it as quickly as possible to intermediary metabolism.
  5. Remove Branches (linearize).
  6. Maximize net energy gain (to CO2) or valuable intermediates.
  7. Minimize leaps of evolution (dissimilarity to known enzymes).
  8. Aerobically, oxido-reductases are common.
  9. Anaerobically, hydrolases are common.

Contact Us
© 2022, EAWAG. All rights reserved.

http://eawag-bbd.ethz.ch/workshop/minutes.html