| 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 | 可制御性 / 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 / 最大マッチング / 複雑生物情報ネットワーク |
|---|
| 研究実績の概要 |
One of the important challenges in computational biology is to control complex biological networks and pathways by only using a few molecules as input information. Network controllability emerged as a new area that combines control theory with network science.
In this research, we integrate heterogeneous biological data (gene expression profiles, and networks and pathways data) with controllability concepts to develop new control-theory based models that can be further applied to analyze and control biological systems.
Because some networks are probabilistic (interactions are not well-determined or random failures may occur), we used a Probabilistic Minimum Dominating set model (PMDS) to identify a few number of controllers that can drive the entire network. Because the solutions are not unique we extended the model and proposed a new Critical PMDS (CPMDS) that efficiently determines critical nodes in dynamically reconstructed probabilistic networks. The model was then applied to investigate time-driven processes such as human aging by integrating a large number of gene expression profiles of samples from different ages. On the other hand, it is well-known that the linear dynamics of the controllability process in networks can also be studied using a maximum matching approach. Our research unveiled a significant difference between the number of driven nodes and that of driver nodes. Moreover, we proposed a novel algorithm to identify critical driven nodes in metabolic pathways. The results highlighted the importance of the analysis of driven nodes in biological networks.
|
| 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
The control analysis of several metabolic fluxes in several human tissues when dynamically transit from normal to cancer states using a probabilistic control model showed that cancer states require fewer controllers than the healthy states. We performed an extension of this model so that currently it allows us to determine critical nodes (e.g. proteins) not only in static networks but also in dynamically constructed networks. We validated the model using temporal process such human aging by collecting gene expression profiles at many different age categories. Efficiency of the new algorithm was also evaluated using exhaustive computer simulations. On the other hand, we proposed a new algorithm to efficiently identify driven nodes in networks. While most of data-driven research has focused on the analysis of driver nodes, the controllability analysis using driven nodes (nodes targeted by signals) captures new insights from biological systems, and may offer a new perspective on network control.
|
| 今後の研究の推進方策 |
The developed critical probabilistic model can be used to analyze not only probabilistic networks but also dynamically constructed networks. In spite of being an NP-hard problem, several uncovered mathematical propositions enhanced the efficiency of the algorithm. However, the algorithm can only address undirected networks. Therefore, we are working on developing a new model that is able to determine critical controllers in probabilistic directed networks. The research involves the mathematical analysis on new suitable propositions as well as extensive computer simulations and data analysis. Moreover, the importance of intermittent nodes (nodes that appear in at least on control set) would also be of research interest. Several computational experiments will also be done to examine/verify control model variations under different control settings.
|
| 次年度使用額が生じた理由 |
今年も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、投稿中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。また、コンピュータシミュレーションを行うため、コンピュータ・ワークステーションを購入する必要がある可能性があります。
|
