蛋白质折叠方式的确定方法、装置、电子设备和存储介质
By constructing a support point pair representation matrix and a support point-residue pair representation matrix, and combining them with a folding pattern prediction model, the problem of low accuracy in protein structure prediction by artificial intelligence algorithms is solved, achieving higher prediction accuracy and autonomy.
CN115798572BActive Publication Date: 2026-07-17SHUIMU FUTURE (HANGZHOU) TECH CO LTD +1
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHUIMU FUTURE (HANGZHOU) TECH CO LTD
- Filing Date
- 2022-11-04
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing artificial intelligence algorithms have low accuracy in predicting protein structures, requiring manual intervention for correction, and thus have significant limitations.
Method used
By determining the amino acid sequence and electron density map of the target protein, support points are screened, and support point-residue pair representation matrices and support point-residue pair representation matrices are constructed. Combined with a folding pattern prediction model, the folding pattern of the protein is predicted.
Benefits of technology
It improves the accuracy of protein folding prediction, reduces reliance on human intervention, and enhances the robustness and generalization ability of the algorithm.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115798572B_ABST
Abstract
本公开涉及计算机和生物学领域，尤其涉及一种蛋白质折叠方式的确定方法、装置、电子设备和存储介质，确定目标蛋白质对应的氨基酸序列以及电子密度图，氨基酸序列中包括至少两个氨基酸。根据电子密度图确定至少一个支撑点，根据每两个支撑点之间的位置关系确定支撑点对表示矩阵，根据支撑点的数量和氨基酸序列中氨基酸数量确定支撑点‑残基对表示矩阵。根据氨基酸序列、支撑点对表示矩阵和支撑点‑残基对表示矩阵确定目标蛋白质的折叠方式。本公开通过根据电子密度图确定其中不同支撑点的空间位置，以及不同支撑点和氨基酸序列中氨基酸的对应关系，与氨基酸序列一同预测目标蛋白质的折叠方式，提高了预测结果的准确性。
Need to check novelty before this filing date? Find Prior Art