Match samples had the distinctive saturate pattern typically found in samples containing petroleum. Probable match samples also had evidence of petroleum saturates in conjunction

with typical background hydrocarbons found in coastal marshes. Samples falling into the match and probable match categories also had evidence of a detectable unresolved complex mixture (UCM) which provides strong evidence of petroleum contamination. Inconclusive and non-match samples mostly contained chromatographic patterns that have been typical of background selleck chemicals hydrocarbons in marsh sediments from Barataria Bay. The diagnostic ratios calculated herein are seemingly robust down to a concentration of ∼200 parts per billion (ppb) of target PAHs (Table 3). Samples with concentrations lower than this contained sufficient levels of background hydrocarbon compounds that interfered with or made impossible the calculation of the ratios, which in turn affected the final sample score. There were non-match samples that had concentrations above the 200 ppb threshold, which provides strong evidence that diagnostic biomarker ratio analysis can distinguish between different

sources of oil in the environment. Even though there were non-match samples with PAH concentrations higher than 200 ppb, overall, low concentrations of oil can introduce error in the calculation of diagnostic ratios. Another important factor to consider is the eventual weathering of the biomarker compounds themselves. click here Biomarker compounds have been shown to be degraded

by severe environmental weathering processes over the course of decades (Wang et al., 2001); however, the degree of degradation depends greatly on sediment organic carbon content, prior exposure to petrogenic hydrocarbons, anoxic and low-energy environmental conditions (Reddy et al., 2002) and whether or not oil residues are buried or remain at the surface. If the biomarkers do indeed weather, this could adversely affect the final categorization of samples; in essence, samples that would have been a match early on during a spill Mannose-binding protein-associated serine protease could end up in the inconclusive or even non-match categories. The conditions above are in no way discouraging the use of diagnostic ratio analysis but are instead given to increase awareness of factors that may limit their effectiveness. Overall, diagnostic ratio analysis and the statistical similarity analysis of inconclusive samples provided a quantifiable and robust categorization of sediment samples. PVA performance was assessed based on two or more vertices’ (or vectors’) normalization or non-normalization of input data, and various methods of identifying vertices. The highest performance was found with a two-vector extreme sample-set solution describing the non-normalized input data (diagnostic ratios) variance.