- Results as:
- Print view
- Endnote (RIS)
- BibTeX
- Table: CSV | HTML
Books and chapters
Books and chapters
Research Mission and Profile
Molecular Signal Processing
Bioorganic Chemistry
Biochemistry of Plant Interactions
Cell and Metabolic Biology
Independent Junior Research Groups
Program Center MetaCom
Publications
Good Scientific Practice
Research Funding
Networks and Collaborative Projects
Symposia and Colloquia
Alumni Research Groups
Books and chapters
The meaningful correlation of sensory data with analytical data is one of the most challenging tasks in flavor research. In beef stocks in particular, due to the presence of low levels of aroma-active compounds and the taste contribution of non-volatile molecules to the typical “juiciness” character, the consumer encounters a complex matrix situation. The goal of our study was to carry out a comprehensive analysis of all relevant flavor molecules and the correlation to human sensory data.A technique recently developed at the IPB and termed “reverse metabolomics” was used to link biological activity (i.e., sensory data) with variations in the metabolic profile (i.e., analytical data). We used this methodology for the first time to correlate sensorial attributes and GC-MS, LC-MS, and NMR data in culinary beef stocks.Reverse metabolomics was applied to study the link between sensory and chemical composition in a series of freshly prepared culinary beef stocks. A set of 10 different beef stocks was prepared. The degree of liking of the samples was recorded on a hedonic 1–9 scale. Analysis of the stocks was performed by LC-MS, GC-MS, and NMR. 1H-NMR data directly obtained from the meat stock were very complex.Analysis of this dataset by reverse metabolomics revealed some basic structural elements of the key taste compounds, such as carnosine or anserine. The reverse metabolomics correlation of quantitative data with partiality revealed the importance of a set of compounds. This relevance of these compounds has been confirmed by additional sensory experiments which showed an increase in perceived juiciness.
Books and chapters
Probabilistic latent semantic indexing (PLSI) represents documents of a collection as mixture proportions of latent topics, which are learned from the collection by an expectation maximization (EM) algorithm. New documents or queries need to be folded into the latent topic space by a simplified version of the EM-algorithm. During PLSI- Folding-in of a new document, the topic mixtures of the known documents are ignored. This may lead to a suboptimal model of the extended collection. Our new approach incorporates the topic mixtures of the known documents in a Bayesian way during folding- in. That knowledge is modeled as prior distribution over the topic simplex using a kernel density estimate of Dirichlet kernels. We demonstrate the advantages of the new Bayesian folding-in using real text data.