In-silico identification of phenotype-biased functional modules

BACKGROUND: Phenotypes exhibited by microorganisms can be useful for several purposes, e.g., ethanol as an alternate fuel. Sometimes, the target phenotype maybe required in combination with other phenotypes, in order to be useful, for e.g., an industrial process may require that the organism survive in an anaerobic, alcohol rich environment and be able to feed on both hexose and pentose sugars to produce ethanol. This combination of traits may not be available in any existing organism or if they do exist, the mechanisms involved in the phenotype-expression may not be efficient enough to be useful. Thus, it may be required to genetically modify microorganisms. However, before any genetic modification can take place, it is important to identify the underlying cellular subsystems responsible for the expression of the target phenotype. RESULTS: In this paper, we develop a method to identify statistically significant and phenotypically-biased functional modules. The method can compare the organismal network information from hundreds of phenotype expressing and phenotype non-expressing organisms to identify cellular subsystems that are more prone to occur in phenotype-expressing organisms than in phenotype non-expressing organisms. We have provided literature evidence that the phenotype-biased modules identified for phenotypes such as hydrogen production (dark and light fermentation), respiration, gram-positive, gram-negative and motility, are indeed phenotype-related. CONCLUSION: Thus we have proposed a methodology to identify phenotype-biased cellular subsystems. We have shown the effectiveness of our methodology by applying it to several target phenotypes. The code and all supplemental files can be downloaded from (http://freescience.org/cs/phenotype-biased-biclusters/).