| 研究課題/領域番号 |
18K11535
|
|---|
| 研究機関 | 東邦大学 |
|---|
研究代表者 |
ホセ ナチェル 東邦大学, 理学部, 教授 (60452984)
|
|---|
| 研究期間 (年度) |
2018-04-01 – 2022-03-31
|
| キーワード | 制御性 / 情報解析 / 代謝経路 / タンパク質相互作用ネットワーク / 遺伝子発現データ / 支配集合 / 複雑生物情報ネットワーク |
|---|
| 研究実績の概要 |
Network controllability aims to integrate network science with control theory in order to control complex networks at will.
Here, we combine heterogeneous data and network controllability concepts with dynamic process/transitions in cell biology. We then focus on proposing methods to control biochemical networks and pathways so that we can investigate how to drive a cell from an abnormal state to a normal state. We used a Probabilistic Minimum Dominating Set (PMDS) model, which identifies a minimum set of nodes. These driver nodes may control the entire network before and after transition. The results indicate that cancer metabolism is characterised by more streamlined flux distributions. The PMDS method is currently being extended to identify efficiently all control categories in dynamically constructed networks, which are probabilistic. Thus, it may be possible to study time-related processes such as human aging. On the other hand, we are also studying the maximum matching controllability approach, and developing new algorithms to identify critical driven nodes. Differences between the number of driver and driven nodes may have relevant biological implications for the applicability of the theory.
Finally, we also proposed a method that uses spectral clustering to analyze complex networks. The method integrates protein networks and gene expression profiles to classify lung cancer using a novel convolutional neural network (CNN) approach. Main novelty is that input data can be networks so it expands the applicability of CNN.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
We have completed an analysis of the control and dynamics of cancer in several human tissues by studying metabolic fluxes in both normal and cancer states. The integration of control theory with flux correlation analysis shows that flux correlations substantially increase in cancer states of breast, kidney and urothelial. Moreover, the PMDS analysis shows that cancer states require fewer controllers than their corresponding healthy states. These results indicate that cancer metabolism is characterised by more streamlined flux distributions. Extensions of the PMDS model are being done to allow us to identify critical proteins in dynamically constructed networks. This analysis may identify critical proteins in longitudinal (temporal) processes such human aging. Moreover, the development of algorithms and data analysis to identify critical driven nodes in metabolic pathways is also very advanced.
We also developed an approach that allows us to analyse biological networks using CNN methods, which is the main novelty of the method. We then integrated protein networks with gene expression profiles for normal and cancer samples. The method was able to classified cancer with high predictive performance in terms of accuracy.
|
| 今後の研究の推進方策 |
We are extending the PMDS approach to enable us to identify not only a PMDS, but also all the control categories, including critical, intermittent and redundant. The PMDS is probabilistic, and relies on a weighted network. The computation is quite intensive. Therefore, we have to efficiently compute by adding some preprocessing steps in terms of mathematical propositions. The method also constructed dynamic networks, so we can investigate the dynamic changes of critical protein fractions along human aging process. We also examine different approach such maximum matching methodology, and whether the differences observed in critical driven and driven nodes have an impact in its biological application. To that end, we have been developing new algorithms as well as performing data analysis. We expect to obtain final results soon.
|
| 次年度使用額が生じた理由 |
今年、2月・３月に出張する予定でしたが、コロナの関係で行けませんでした。
次年度以降は、準備中の2つの論文と今後執筆予定の複数の新しい論文の出版料及び別刷り料が必要となる可能性があります。パンデミックの状況による国内発表・出張なども行う可能性あります。
|
