Search: WFRF:(Brunius Carl 1974)
> (2022) >
Identifying metabot...
Identifying metabotypes from tensor data
-
- Skantze, Viktor, 1992 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Wallman, Mikael, 1979 (author)
- Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik (FCC),Fraunhofer-Chalmers Research Centre for Industrial Mathematics (FCC)
-
- Sandberg, Ann-Sofie, 1951 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
show more...
-
- Landberg, Rikard, 1981 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Jirstrand, Mats, 1968 (author)
- Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik (FCC),Fraunhofer-Chalmers Research Centre for Industrial Mathematics (FCC)
-
- Brunius, Carl, 1974 (author)
- Chalmers tekniska högskola,Chalmers University of Technology
-
show less...
-
(creator_code:org_t)
- 2022
- 2022
- English.
- Related links:
-
https://research.cha...
Abstract
Subject headings
Close
- Metabolic response to diet shows large individual variation, which warrants tailored dietary recommendation i.e., personalized nutrition (PN). A step towards PN is to tailor diet to groups of individuals with similar metabolic phenotype, so called metabotypes (i.e., clusters of individuals with similar metabolism). Metabotyping of high-dimensional data is commonly performed in matrix form using matrix decompositions (e.g., PCA). However, data from e.g., crossover studies can be conveniently organized in multi-dimensional form (i.e., as tensor data) and methods for detecting metabotypes in such data are still lacking. We therefore aimed to develop and evaluate tools to identify potential metabotypes in high-dimensional tensor data. We developed two methods: The first uses CANDECOMP/PARAFAC (CP) decomposition directly on tensor data where clustering was performed on individual’s scores, whereas the second was developed specifically for time-resolved data and uses dynamic mode decomposition (DMD) to model metabolite dynamics, where clustering was performed on individual’s dynamic state trajectories. We applied the methods to identify metabotypes in data from a crossover acute post-prandial dietary intervention study on 17 overweight males (BMI 25–30 kg/m2, 41–67 y of age) undergoing three dietary interventions (pickled herring, baked herring and baked beef, measuring 79 metabolites (from GC-MS metabolomics) at 8 time points (0-7h). Both methods identified two potential metabotype clusters, predominantly in amino acids after the meat diet. The clustering associated to baseline levels of creatinine, strengthening the plausibility of found metabotypes. The CP method is a general approach, not specific to time-resolved data, and provides better fit if the data is multilinear. Conversely, DMD is designed for time-resolved data, for which it often provides a better fit than CP. We concluded that both the CP and the DMD approach are well suited to identify metabotypes in tensor data from a wide variety of complex experimental designs.
Subject headings
- NATURVETENSKAP -- Data- och informationsvetenskap -- Annan data- och informationsvetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Other Computer and Information Science (hsv//eng)
- NATURVETENSKAP -- Biologi -- Bioinformatik och systembiologi (hsv//swe)
- NATURAL SCIENCES -- Biological Sciences -- Bioinformatics and Systems Biology (hsv//eng)
- MEDICIN OCH HÄLSOVETENSKAP -- Hälsovetenskap -- Näringslära (hsv//swe)
- MEDICAL AND HEALTH SCIENCES -- Health Sciences -- Nutrition and Dietetics (hsv//eng)
Keyword
- Metabolomics
- PARAFAC
- Tensor decomposition
- Dynamic Mode Decomposition
Publication and Content Type
- kon (subject category)
- vet (subject category)
To the university's database
- By the author/editor
-
Skantze, Viktor, ...
-
Wallman, Mikael, ...
-
Sandberg, Ann-So ...
-
Landberg, Rikard ...
-
Jirstrand, Mats, ...
-
Brunius, Carl, 1 ...
- About the subject
-
- NATURAL SCIENCES
-
NATURAL SCIENCES
-
and Computer and Inf ...
-
and Other Computer a ...
-
- NATURAL SCIENCES
-
NATURAL SCIENCES
-
and Biological Scien ...
-
and Bioinformatics a ...
-
- MEDICAL AND HEALTH SCIENCES
-
MEDICAL AND HEAL ...
-
and Health Sciences
-
and Nutrition and Di ...
- By the university
-
Chalmers University of Technology