FIGURE 1.

Demonstration of the limitations of sequence homology based pathway prediction algorithms. In this figure the hypothetical pathway ‘A–F’ comprises of six genes representing different candidate bacterial genomes Genome 1, 2, 3, 4. The figure illustrates four cases of gene arrangement where Genomes 1 and 2 do not follow the operon model as they have all the constituent six genes but dispersed at distant locations on the genome, Genome 3 has multiple homologs of the constituent genes dispersed on the genome. These dispersed homologs may function in pathways other than A–F. Genome 4 consists of all the six genes lying within context and represents the actual pathway. Homology based methods might inaccurately predict pathway presence in all the four cases. The actual presence of pathway is observed only in Genome 4 which can be correctly captured by gene context based method.