2018 Fiscal Year Research-status Report
複雑生物情報ネットワークにおけるダイナミクスと制御性の統合情報解析
| Project/Area Number |
18K11535
|
|---|
| Research Institution | Toho University |
|---|
Principal Investigator |
ホセ ナチェル 東邦大学, 理学部, 教授 (60452984)
|
|---|
| Project Period (FY) |
2018-04-01 – 2022-03-31
|
| Keywords | 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 |
|---|
| Outline of Annual Research Achievements |
The complete understanding of a complex system is measured by our ability to control it. In a cell, biological networks govern complex cellular processes as well as diseases development. We, therefore, aim to develop computational methods that allow us to control biochemical pathways such that we can actually drive a cell from an abnormal state to a normal state. We then focus on integrating controllability methods with multiple types of biological data, including gene expression profiles for different diseases states and biological processes.
We have considered complex structures defined in terms of multilayer networks and developed algorithms to simultaneously identify a set of molecules that control all networks. The dynamics of cancer in several human tissues as well as viral infection HIV-1 and hepatities C virus have also been investigated by proposing dynamic models that identify a set of controllers responsible for the underlying process.
Several methods on integration of biological data were also investigated including non-linear gene regulation model driven by gene expression data as well as combining protein networks and gene expression profiles to classify cancer states using a novel convolutional neural network approach.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
We defined the multilayer control problem in terms of the minimum dominating set (MDS) controllability framework. We then devised a fast algorithm that can determine the minimum set of enzymes necessary to control simultaneously multilayer metabolic networks in plants.
Dynamics of viral infection is also a central topic in computational biology. Recently, we used viral infection, specifically human immunodeficiency virus type 1 (HIV-1) and hepatitis C virus (HCV), as a paradigm to model control of an infected cell and proposed an algorithmic technique to identify viral control proteins.
We also proposed a new convolutional neural network approach to cancer classification by integrating protein interaction network data and gene expression profiles. On the other hand, We have also examined the dynamics of cancer in several human tissues by reconstructing metabolic fluxes in both normal and cancer states. Preliminary results indicate that controllability analysis aids to elucidate how cancer reshapes the human metabolism.
|
| Strategy for Future Research Activity |
Based on our research results, the combination of probabilistic controllability methods with metabolic fluxes data at different states may aid to elucidate the transition from normal to cancer state. We expect to extend this research direction.
By using similar probabilistic controllability techniques, we aim to classify the controllers based on their control roles. Our theoretical analysis has derived three theorems that may enhance computational time so that we can also identify control categories even in a computational expensive method such as the proposed probabilistic control model. We can then apply it to understand the aging process in human by identifying key control proteins. Dynamic construction of temporal networks as well as gene expression profiles at different age would be used in the study.
In parallel, we also plan to evaluate and validate the developed dynamic mathematical models -in terms of differential equations- that can be integrated with gene expression profiles to determine molecular controllers in different diseases.
|
| Causes of Carryover |
今年、私は3つの国際会議で講演するように招待されました。さらに、2つの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。そこで、私は次の年度に、残りの予算を移動することに決めました。
|
