Processing ethanol from cellulose -- wheat and rice straw, switchgrass, paper pulp, agricultural waste products like corn cobs and leaves -- has the potential to squeeze at least twice as much fuel from the same area of land, because so much more biomass is available per acre. Moreover, such an approach would use feedstocks that are otherwise essentially worthless.
Converting cellulose to ethanol involves two fundamental steps: breaking the long chains of cellulose molecules into glucose and other sugars, and fermenting those sugars into ethanol. In nature, these processes are performed by different organisms: fungi and bacteria that use enzymes (cellulases) to "free" the sugar in cellulose, and other microbes, primarily yeasts, that ferment sugars into alcohol.
The ideal organism would do it all -- break down cellulose like a bacterium, ferment sugar like a yeast, tolerate high concentrations of ethanol, and devote most of its metabolic resources to producing just ethanol. There are two strategies for creating such an all-purpose bug. One is to modify an existing microbe by adding desired genetic pathways from other organisms and "knocking out" undesirable ones; the other is to start with the clean slate of a stripped-down synthetic cell and build a custom genome almost from scratch. Synthetic Genomics, founded by Craig Venter, is in hot pursuit of a bacterium. There is progress in both strategies.