Finding Drug Targets with Flux Balance Analysis

Flux balance analysis (FBA) is an in silico metabolic analysis technique that has been used to identify potential drug targets in pathogens (disease-causing organisms) by determining essential biochemical reactions in these organisms [1]. To do this, previous studies have generally focused on completely removing each reaction in turn and checking if the pathogens can still survive, without quantifying the extent to which a reaction needs to be knocked out in order to be fatal. Thus, previous analysis relies on being able to completely knock down genes that code for the enzymes (or in some other way disable the enzymes) that catalyze certain reactions, which is often hard to do in reality; instead, it is more common for a drug to only partially decrease the rate of a reaction. Moreover, in a typical cell, metabolic networks are highly redundant: if the enzyme that catalyzes one reaction is knocked out, other reactions in the cell can become more active to compensate. Thus, instead of focusing on only single reactions, it is desirable to consider drugs that can act on several reactions at once. In this project, we extend prior work to deal with these issues of partial knockdowns and multiple targets, allowing us to better identify potential targets for drugs that are e ective in killing pathogens. We demonstrate the e ectiveness of our approach on synthetic metabolic pathway data as well as real-world metabolic data from the tuberculosis-causing bacterium, Mycobacterium tuberculosis.