WEKO3
アイテム
アミノ酸配列情報のみからの深層学習によるヒト-ウイルスタンパク質間相互作用の予測に関する研究
https://doi.org/10.18997/0002000934
https://doi.org/10.18997/000200093477db3ab8-8480-4a0c-b031-532f7b931487
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
jou_k_396.pdf (4.9 MB)
|
|
| アイテムタイプ | 学位論文 = Thesis or Dissertation(1) | |||||||
|---|---|---|---|---|---|---|---|---|
| 公開日 | 2024-08-28 | |||||||
| 資源タイプ | ||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_db06 | |||||||
| 資源タイプ | doctoral thesis | |||||||
| タイトル | ||||||||
| タイトル | Sequence-based human-virus protein-protein interaction prediction using deep learning | |||||||
| 言語 | en | |||||||
| タイトル | ||||||||
| タイトル | アミノ酸配列情報のみからの深層学習によるヒト-ウイルスタンパク質間相互作用の予測に関する研究 | |||||||
| 言語 | ja | |||||||
| 言語 | ||||||||
| 言語 | eng | |||||||
| 著者 |
築山, 翔
× 築山, 翔
|
|||||||
| 抄録 | ||||||||
| 内容記述タイプ | Abstract | |||||||
| 内容記述 | Viral infection is a significant global health concern. SARS-CoV-2 rapidly spread from its emergence and caused a worldwide pandemic. After invading host cells through attachments with the receptors on the host cell’s surface, viruses hijack host cell functions, such as the cell cycle and apoptosis, by incorporating their genetic material. Because protein-protein interactions (PPIs) play a crucial role in those processes, identifying interactions between human and virus proteins is important for understanding the mechanisms of viral infection and discovering effective drugs and vaccines. Experimental methods, including mass spectrometry and yeast two-hybrid assays, are widely used for identifying human-virus PPI (HV-PPI), but they are time-consuming and costly. Therefore, it is not practical to apply it to all protein pairs. To complement the experimental methods, computational methods have been developed. While several protein structure-based methods, such as molecular dynamics simulation, are able to predict interactions with high accuracy, identifying the structure information is not an easy task, which limits its applicability for proteins with unknown structures. Thereby, recently, PPI prediction methods based on amino acid sequence alone have been introduced. While those methods are receiving a lot of attention, there is room for improvement in the prediction performances, especially novel viruses. To overcome this issue, in the present study, we developed two deep learning-based HV-PPI predictors: LSTM-PHV and Cross-Attention PHV. In those workflows, amino acid sequences were transformed feature representations by word2vec, and HV-PPIs were predicted by neural network models. Both models presented accuracies > 0.95 and AUCs > 0.97 in benchmark datasets and achieved higher prediction accuracy than the existing methods. These methods demonstrated superior performance not only in well-known virus species but also in novel virus species. In addition, the Cross-Attention PHV enables the PPI prediction of proteins with a long length, exceeding 9000 in length. These methods present a substantial advancement over current approaches, addressing their limitations and promising to extend the capabilities of sequencebased PPI prediction methods. This enhancement is expected to contribute significantly to the understanding of viral infection mechanisms and the development of novel antiviral drugs. To facilitate the research fields, we implemented web servers which are accessible from http://kurata35.bio.kyutech.ac.jp/LSTM-PHV/ and http://kurata35.bio.kyutech.ac.jp/Crossattention_ PHV/ in LSTM-PHV and Cross-Attention PHV, respectively. |
|||||||
| 目次 | ||||||||
| 内容記述タイプ | TableOfContents | |||||||
| 内容記述 | 1.Introduction| 2. Sequence-based PPI prediction| 3.LSTM-PHV| 4. Cross-Attention PHV| 5.Conclusion| 6. Future works | |||||||
| 備考 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 九州工業大学博士学位論文 学位記番号: 情工博甲第396号 学位授与年月日:令和6年3月25日 | |||||||
| 学位授与番号 | ||||||||
| 学位授与番号 | 甲第396号 | |||||||
| 学位名 | ||||||||
| 学位名 | 博士（情報工学） | |||||||
| 学位授与年月日 | ||||||||
| 学位授与年月日 | 2024-03-25 | |||||||
| 学位授与機関 | ||||||||
| 学位授与機関識別子Scheme | kakenhi | |||||||
| 学位授与機関識別子 | 17104 | |||||||
| 学位授与機関名 | 九州工業大学 | |||||||
| 言語 | ja | |||||||
| 学位授与年度 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 令和5年度 | |||||||
| 出版タイプ | ||||||||
| 出版タイプ | VoR | |||||||
| 出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 | |||||||
| アクセス権 | ||||||||
| アクセス権 | open access | |||||||
| アクセス権URI | http://purl.org/coar/access_right/c_abf2 | |||||||
| ID登録 | ||||||||
| ID登録 | 10.18997/0002000934 | |||||||
| ID登録タイプ | JaLC | |||||||
