Establishing parameter estimation theory of stochastic differential equations for advanced modeling of life systems

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K12059
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 62010:Life, health and medical informatics-related
• Research Institution: The University of Tokyo
• Principal Investigator: Kiryu Hisanori 東京大学, 大学院新領域創成科学研究科, 准教授 (80415778)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 一細胞シーケンシング / 機械学習 / 確率モデル / トランスクリプトーム / 確率偏微分方程式

Outline of Final Research Achievements

Spatio-temporal information on biological processes is rapidly increasing due to DNA sequencing technology and camera performance improvements. This has made it possible to rigorously study the temporal causality of gene-gene interactions and the effects of the three-dimensional arrangement of cells and tissues on the behavior of organisms. Therefore, this study aimed to develop and implement a general-purpose machine learning technique to estimate parameters of nonlinear stochastic partial differential equations from data as a tool to enable more sophisticated modeling of life processes. We tried various methods, such as computer algebra, automatic differentiation, and diffusion models. However, many issues remain to be solved to make it a generally usable tool, and further research should be continued.

Translated with www.DeepL.com/Translator (free version)

Free Research Field

バイオインフォマティクス

Academic Significance and Societal Importance of the Research Achievements

近年、生物に関する細胞レベルの非常に高い解像度の時空間的情報が爆発的に蓄積しており、これらを有効に活用して、未知の生物機構を発見したり、病気の治療に役立つ情報を抽出する情報科学的手法の開発の重要性が増している。本研究では確率偏微分方程式のパラメータ推定論に取り組んだが、これに成功すれば、遺伝子発現情報などから未来の生物の変化を予測するための強力なツールとなることが期待できる。