Improving Phylogenetic Tree Accuracy through Recommending Missing Metabolic pathways: A Collaborative Filtering Approach
Lists
Only you see this
Abstract
With the accumulation of massive biological datasets, the constructions of phylogenetic trees become versatile. For instance, metabolomics data accommodate profiles of all metabolites and metabolic pathways in species, and thus such information can be used for phylogenetic tree construction. The basic assumption of using metabolic pathways to construct phylogenetic trees is that phylogenetically closer species should have more similar pathways and thus they should be clustered together. One of the problems of using metabolic pathways for tree construction is that some metabolic pathway dataset might be missing in one species or another. In this paper, we proposed a Collaborative Filtering (CF) based recommender system to fill up the missing metabolic pathways, and dataset normalization to better the tree accuracy. Our experimental results on both prokaryotic and eukaryotic phylogenetic trees have shown that our approach could improve clustering accuracy. Therefore, we conclude that our collaborative filtering recommender system is useful for improving the phylogenetic tree constructed with metabolic pathways.
BibTeX
@inproceedings{wan2014improving,
author = {Peng Wan and Dongsheng Che},
title = {{Improving Phylogenetic Tree Accuracy through Recommending Missing Metabolic pathways: A Collaborative Filtering Approach}},
booktitle = {American},
year = {2014},
doi = {10.7726/ajbcb.2014.1002},
}Loading related content...