Nonnegative Matrix Factorization for Dynamic Modules in Cancer Attribute Temporal Networks
Lists
Only you see this
Abstract
Tracking the dynamic modules of genome associated with cancer progression is essential for studying cancer pathogenesis, diagnosis and treatment. The current algorithms only focus on extracting dynamic modules from temporal networks associated with cancer progression, where networks are homogeneous. However, as far as we know, there is no attempt has been made to extract dynamic modules through fusing gene attribute information with gene co-expression temporal networks. To solve this problem, an novel algorithm (aka TANMF) is proposed to detect dynamic epigenetic modules in cancer attribute networks, which is formulated as an optimization problem. Specifically, we decompose the gene attribute matrix and gene co-expression temporal networks under the same basis matrix at each stage, and sparse the feature matrix by L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> regularization to increase the robustness of the algorithm. Experimental results demonstrate that the proposed algorithm is more accurate than state-of-the-art methods on simulated data. By applying TANMF to breast cancer data, the obtained dynamic modules are more enriched by the known pathways. The proposed model and algorithm provide an effective way to introduce gene attribute to investigate dynamics of cancer.
BibTeX
@inproceedings{li2019nonnegative,
author = {Dongyuan Li and Xiaoke Ma},
title = {{Nonnegative Matrix Factorization for Dynamic Modules in Cancer Attribute Temporal Networks}},
booktitle = {BIBM},
year = {2019},
doi = {10.1109/bibm47256.2019.8983045},
}Loading related content...