Limits...
A weighted relative difference accumulation algorithm for dynamic metabolomics data: long-term elevated bile acids are risk factors for hepatocellular carcinoma.

Zhang W, Zhou L, Yin P, Wang J, Lu X, Wang X, Chen J, Lin X, Xu G - Sci Rep (2015)

Bottom Line: To extract important information in a systematic time dimension rather than at isolated time points, a weighted method based on the means and variations along the time points was proposed and first applied to previously published rat model data.Permutation was employed for noise filtering and false discovery rate (FDR) was used for parameter optimization during the feature selection.Long-term elevated serum bile acids were identified as risk factors for HCC development.

View Article: PubMed Central - PubMed

Affiliation: School of Computer Science &Technology, Dalian University of Technology, Dalian, China.

ABSTRACT
Dynamic metabolomics studies can provide a systematic view of the metabolic trajectory during disease development and drug treatment and reveal the nature of biological processes at metabolic level. To extract important information in a systematic time dimension rather than at isolated time points, a weighted method based on the means and variations along the time points was proposed and first applied to previously published rat model data. The method was subsequently extended and applied to prospective metabolomics data analysis of hepatocellular carcinoma (HCC). Permutation was employed for noise filtering and false discovery rate (FDR) was used for parameter optimization during the feature selection. Long-term elevated serum bile acids were identified as risk factors for HCC development.

No MeSH data available.


Related in: MedlinePlus

Flow chart of the analysis of the rat metabolomics data.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4355672&req=5

f1: Flow chart of the analysis of the rat metabolomics data.

Mentions: Parallel to our previous feature-defining process, a two-level data analysis procedure employing the wRDA (Fig. 1) was performed to select meaningful features to discriminate between the models and control, and between HCC and non-HCC samples. In the first level, 152 ion features were removed by means of permutation, leaving 1092 features with a false discovery rate (FDR) of 0 to constitute feature subset 1. Then, Support Vector Machine36 (SVM) was conducted based on the top 20 features (Fig. 2A) ranked by the wRDA. Five fold cross validation was conducted 50 times. The average accuracy rate was 98.85% ± 0.66%, demonstrating that the top ranked variables have a strong ability to distinguish disease samples from controls. These metabolic features include two bile acids (TCA and tauroursodeoxycholic acid (TUDCA)), LPCs, and LPEs with different acyl chains and unsaturation levels. These results indicate a disturbance of lipid metabolism in DEN-induced liver disease.


A weighted relative difference accumulation algorithm for dynamic metabolomics data: long-term elevated bile acids are risk factors for hepatocellular carcinoma.

Zhang W, Zhou L, Yin P, Wang J, Lu X, Wang X, Chen J, Lin X, Xu G - Sci Rep (2015)

Flow chart of the analysis of the rat metabolomics data.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4355672&req=5

f1: Flow chart of the analysis of the rat metabolomics data.
Mentions: Parallel to our previous feature-defining process, a two-level data analysis procedure employing the wRDA (Fig. 1) was performed to select meaningful features to discriminate between the models and control, and between HCC and non-HCC samples. In the first level, 152 ion features were removed by means of permutation, leaving 1092 features with a false discovery rate (FDR) of 0 to constitute feature subset 1. Then, Support Vector Machine36 (SVM) was conducted based on the top 20 features (Fig. 2A) ranked by the wRDA. Five fold cross validation was conducted 50 times. The average accuracy rate was 98.85% ± 0.66%, demonstrating that the top ranked variables have a strong ability to distinguish disease samples from controls. These metabolic features include two bile acids (TCA and tauroursodeoxycholic acid (TUDCA)), LPCs, and LPEs with different acyl chains and unsaturation levels. These results indicate a disturbance of lipid metabolism in DEN-induced liver disease.

Bottom Line: To extract important information in a systematic time dimension rather than at isolated time points, a weighted method based on the means and variations along the time points was proposed and first applied to previously published rat model data.Permutation was employed for noise filtering and false discovery rate (FDR) was used for parameter optimization during the feature selection.Long-term elevated serum bile acids were identified as risk factors for HCC development.

View Article: PubMed Central - PubMed

Affiliation: School of Computer Science &Technology, Dalian University of Technology, Dalian, China.

ABSTRACT
Dynamic metabolomics studies can provide a systematic view of the metabolic trajectory during disease development and drug treatment and reveal the nature of biological processes at metabolic level. To extract important information in a systematic time dimension rather than at isolated time points, a weighted method based on the means and variations along the time points was proposed and first applied to previously published rat model data. The method was subsequently extended and applied to prospective metabolomics data analysis of hepatocellular carcinoma (HCC). Permutation was employed for noise filtering and false discovery rate (FDR) was used for parameter optimization during the feature selection. Long-term elevated serum bile acids were identified as risk factors for HCC development.

No MeSH data available.


Related in: MedlinePlus