Improving the accuracy of RNA secondary structure prediction by machine learning based on next-generation sequencing data

2019 Fiscal Year Final Research Report

• Project/Area Number: 16K00404
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Life / Health / Medical informatics
• Research Institution: Keio University
• Principal Investigator: Sato Kengo 慶應義塾大学, 理工学部(矢上), 講師 (20365472)
• Project Period (FY): 2016-04-01 – 2020-03-31
• Keywords: バイオインフォマティクス / RNA二次構造予測 / 機械学習

Outline of Final Research Achievements

We have developed a machine learning algorithm that makes it possible to use secondary structure profiles, which are partial structural information, as weak-level learning data, and aims to improve the accuracy of RNA secondary structure prediction without overfitting by learning a large number of secondary structure models that are more precise than existing methods. First, we developed a more robust and accurate method for RNA secondary structure prediction by integrating the free energy minimization method based on the existing Turner thermodynamic model with the machine learning method using a structured SVM. The results of the computer experiments showed that no overfitting was observed, unlike in the existing methods, and the prediction accuracy was improved.

Free Research Field

バイオインフォマティクス

Academic Significance and Societal Importance of the Research Achievements

RNA 二次構造予測は古くから研究されているが，長い配列に対する予測精度は未だに十分とは言えない．本研究によりRNA二次構造予測の精度が向上し，生体内におけるRNAの機能を推定する手がかりが得られることが期待される．さらに，RNAウィルスをターゲットとする創薬などに応用することが可能である．