Multimodal deep learning for allergenic proteins prediction
Abstract Background Accurate prediction of allergens is essential for identifying the sources of allergic reactions and preventing future exposure to harmful triggers; however, the limited performance of current prediction tools hinders their practical applications. Results Here, we present Multimod...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-07-01
|
| Series: | BMC Biology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12915-025-02347-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background Accurate prediction of allergens is essential for identifying the sources of allergic reactions and preventing future exposure to harmful triggers; however, the limited performance of current prediction tools hinders their practical applications. Results Here, we present Multimodal-AlgPro, a unified framework based on a multimodal deep learning algorithm designed to predict allergens by integrating multiple dimensions, including physicochemical properties, amino acid sequences, and evolutionary information. An exhaustive search strategy for model combinations has also been introduced to ensure robust allergen prediction by thoroughly exploring every possible modality configuration to determine the most effective framework architecture. Additionally, identifying explainable sequence motifs and molecular descriptors from these models that facilitate epitope discovery is of interest. Because it leverages diverse heterogeneous features and our improved multimodal data fusion, Multimodal-AlgPro outperformed several existing methods, demonstrating its potential to significantly advance the accuracy of allergen prediction. Conclusions Overall, Multimodal-AlgPro is a valuable tool for deciphering the mechanisms of allergic responses and offers novel insights on epitope design, with applications in both public health and industrial sectors. |
|---|---|
| ISSN: | 1741-7007 |