In this study preprocessing of Raman spectra of different biological samples has been studied, and their effect on the ability to extract robust and quantitative information has been evaluated. Four data sets of Raman spectra were chosen in order to cover different aspects of biological Raman spectra, and the samples constituted salmon oils, juice samples, salmon meat, and mixtures of fat, protein, and water. A range of frequently used preprocessing methods, as well as combinations of different methods, was evaluated. Different aspects of regression results obtained from partial least squares regression (PLSR) were used as indicators for comparing the effect of different preprocessing methods. The results, as expected, suggest that baseline correction methods should be performed in advance of normalization methods. By performing total intensity normalization after adequate baseline correction, robust calibration models were obtained for all data sets. Combination methods like standard normal variate (SNV), multiplicative signal correction (MSC), and extended multiplicative signal correction (EMSC) in their basic form were not able to handle the baseline features present in several of the data sets, and these methods thus provide no additional benefits compared to the approach of baseline correction in advance of total intensity normalization. EMSC provides additional possibilities that require further investigation.

No abstract

BackgroundOleaginous fungi can accumulate lipids by utilizing a wide range of waste substrates. They are an important source for the industrial production of omega-6 polyunsaturated fatty acids (gamma-linolenic and arachidonic acid) and have been suggested as an alternative route for biodiesel production. Initial research steps for various applications include the screening of fungi in order to find efficient fungal producers with desired fatty acid composition. Traditional cultivation methods (shake flask) and lipid analysis (extraction-gas chromatography) are not applicable for large-scale screening due to their low throughput and time-consuming analysis. Here we present a microcultivation system combined with high-throughput Fourier transform infrared (FTIR) spectroscopy for efficient screening of oleaginous fungi.ResultsThe microcultivation system enables highly reproducible fungal fermentations throughout 12 days of cultivation. Reproducibility was validated by FTIR and HPLC data. Analysis of FTIR spectral ester carbonyl peaks of fungal biomass offered a reliable high-throughput at-line method to monitor lipid accumulation. Partial least square regression between gas chromatography fatty acid data and corresponding FTIR spectral data was used to set up calibration models for the prediction of saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, unsaturation index, total lipid content and main individual fatty acids. High coefficients of determination (R2 = 0.86–0.96) and satisfactory residual predictive deviation of cross-validation (RPDCV = 2.6–5.1) values demonstrated the goodness of these models.ConclusionsWe have demonstrated in this study, that the presented microcultivation system combined with rapid, high-throughput FTIR spectroscopy is a suitable screening platform for oleaginous fungi. Sample preparation for FTIR measurements can be automated to further increase throughput of the system.Electronic supplementary materialThe online version of this article (doi:10.1186/s12934-017-0716-7) contains supplementary material, which is available to authorized users.

a Model-based pre-processing has become wide spread in spectroscopy and is the standard procedure in Fourier-transform infrared spectroscopy. It has also been shown to give valuable contributions in Raman spectroscopy. Extended multiplicative signal correction is flexible enough to handle varying fluorescence background and take into account individual variations in baselines while still keeping enough rigidity through reference spectra and model fitting to avoid degenerate solutions and overfitting, when used correctly. We demonstrate the basic extended multiplicative signal correction method and some extensions, including a novel shift correction, on real Raman data to demonstrate effects on visual appearance, replicate variation and prediction. Comparisons with other standard correction methods are also shown and discussed.