6-21: Identification of biomass fermentation inhibitors using chemometric modeling of HPLC-UV data

Fermentation of lignocellulosic biomass to produce fuel-ethanol holds a broad range of economic and environmental benefits. However, during most pretreatments of biomass (a near-universal prerequisite to fermentation), lignin and various carbohydrates form degradation products that are known to be inhibitory, or that are potentially inhibitory, to downstream fermentation processes. The toxicity of these compounds can be due to both their relative abundance in the sample and to synergistic effects between different sample components. The number of degradation products that have been identified to-date likely represents only a fraction of the components present in pretreatment hydrolysates.  Thus, a continuing need exists for methodologies that promote identification of the sample constituent(s) responsible for observed inhibitory effects in bioconversion processes.

Here, we demonstrate a novel approach for identification of fermentation inhibitors based on chemometric modeling of chromatographic data. The chemometric model was developed using biomass hydrolysates of variable composition, obtained from twenty different dilute-acid pretreatment conditions. Each hydrolysate sample was analyzed via reversed-phase HPLC with UV detection at four discreet wavelengths. Fermentability of each hydrolysate was determined independently in batch fermentation experiments. Chromatographic data were regressed against fermentability data to construct a model, which was subsequently used to identify retention times that were most strongly correlated with observed variations in hydrolysate fermentability.  This approach has strong potential to expedite identification of inhibitory compounds in pretreatment samples.