Mining metabolic networks for optimal drug targets

Area of Science:

• Bioinformatics
• Systems Biology
• Computational Biology

Background:

• Drug design increasingly utilizes bioinformatics to minimize adverse effects.
• Identifying optimal enzyme combinations as drug targets is crucial for efficacy and safety.
• Computational methods are needed to navigate complex metabolic networks for target identification.

Purpose of the Study:

• To formulate and solve the optimal enzyme-combination identification problem within metabolic networks.
• To develop an efficient computational approach for minimizing drug side effects.
• To identify enzyme targets that effectively eliminate target compounds with minimal collateral damage.

Main Methods:

• Formulation of the problem as an optimization task on metabolic networks.
• Development of a graph-based computational damage model to assess enzyme impact.
• Implementation of a branch-and-bound algorithm (OPMET) with pruning strategies for efficient search space exploration.

Main Results:

• OPMET accurately identifies target enzymes for known drugs in the human metabolic network.
• The algorithm demonstrates significant reductions in search time (several orders of magnitude) compared to exhaustive search.
• Experimental validation confirms the efficacy of OPMET in finding optimal enzyme combinations.

Conclusions:

• OPMET provides an efficient and accurate computational solution for identifying optimal enzyme combinations in drug design.
• The developed method effectively balances therapeutic efficacy with the minimization of side effects.
• This approach advances the field of computational drug discovery by optimizing target selection in metabolic networks.