Bacterial Immunogenicity Prediction by Machine Learning Methods
The identification of protective immunogens is the most important and vigorous initial step in the long-lasting and expensive process of vaccine design and development. Machine learning (ML) methods are very effective in data mining and in the analysis of big data such as microbial proteomes. They a...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-11-01
|
| Series: | Vaccines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-393X/8/4/709 |
| _version_ | 1797546261717450752 |
|---|---|
| author | Ivan Dimitrov Nevena Zaharieva Irini Doytchinova |
| author_facet | Ivan Dimitrov Nevena Zaharieva Irini Doytchinova |
| author_sort | Ivan Dimitrov |
| collection | DOAJ |
| description | The identification of protective immunogens is the most important and vigorous initial step in the long-lasting and expensive process of vaccine design and development. Machine learning (ML) methods are very effective in data mining and in the analysis of big data such as microbial proteomes. They are able to significantly reduce the experimental work for discovering novel vaccine candidates. Here, we applied six supervised ML methods (partial least squares-based discriminant analysis, <i>k</i> nearest neighbor (<i>k</i>NN), random forest (RF), support vector machine (SVM), random subspace method (RSM), and extreme gradient boosting) on a set of 317 known bacterial immunogens and 317 bacterial non-immunogens and derived models for immunogenicity prediction. The models were validated by internal cross-validation in 10 groups from the training set and by the external test set. All of them showed good predictive ability, but the xgboost model displays the most prominent ability to identify immunogens by recognizing 84% of the known immunogens in the test set. The combined RSM-<i>k</i>NN model was the best in the recognition of non-immunogens, identifying 92% of them in the test set. The three best performing ML models (xgboost, RSM-<i>k</i>NN, and RF) were implemented in the new version of the server VaxiJen, and the prediction of bacterial immunogens is now based on majority voting. |
| first_indexed | 2024-03-10T14:27:32Z |
| format | Article |
| id | doaj.art-c8bb44b963e8459ca46b1ca615387793 |
| institution | Directory Open Access Journal |
| issn | 2076-393X |
| language | English |
| last_indexed | 2024-03-10T14:27:32Z |
| publishDate | 2020-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Vaccines |
| spelling | doaj.art-c8bb44b963e8459ca46b1ca6153877932023-11-20T22:54:48ZengMDPI AGVaccines2076-393X2020-11-018470910.3390/vaccines8040709Bacterial Immunogenicity Prediction by Machine Learning MethodsIvan Dimitrov0Nevena Zaharieva1Irini Doytchinova2Faculty of Pharmacy, Medical University of Sofia, 1000 Sofia, BulgariaFaculty of Pharmacy, Medical University of Sofia, 1000 Sofia, BulgariaFaculty of Pharmacy, Medical University of Sofia, 1000 Sofia, BulgariaThe identification of protective immunogens is the most important and vigorous initial step in the long-lasting and expensive process of vaccine design and development. Machine learning (ML) methods are very effective in data mining and in the analysis of big data such as microbial proteomes. They are able to significantly reduce the experimental work for discovering novel vaccine candidates. Here, we applied six supervised ML methods (partial least squares-based discriminant analysis, <i>k</i> nearest neighbor (<i>k</i>NN), random forest (RF), support vector machine (SVM), random subspace method (RSM), and extreme gradient boosting) on a set of 317 known bacterial immunogens and 317 bacterial non-immunogens and derived models for immunogenicity prediction. The models were validated by internal cross-validation in 10 groups from the training set and by the external test set. All of them showed good predictive ability, but the xgboost model displays the most prominent ability to identify immunogens by recognizing 84% of the known immunogens in the test set. The combined RSM-<i>k</i>NN model was the best in the recognition of non-immunogens, identifying 92% of them in the test set. The three best performing ML models (xgboost, RSM-<i>k</i>NN, and RF) were implemented in the new version of the server VaxiJen, and the prediction of bacterial immunogens is now based on majority voting.https://www.mdpi.com/2076-393X/8/4/709protective immunogensmachine learningimmunogenicity prediction |
| spellingShingle | Ivan Dimitrov Nevena Zaharieva Irini Doytchinova Bacterial Immunogenicity Prediction by Machine Learning Methods Vaccines protective immunogens machine learning immunogenicity prediction |
| title | Bacterial Immunogenicity Prediction by Machine Learning Methods |
| title_full | Bacterial Immunogenicity Prediction by Machine Learning Methods |
| title_fullStr | Bacterial Immunogenicity Prediction by Machine Learning Methods |
| title_full_unstemmed | Bacterial Immunogenicity Prediction by Machine Learning Methods |
| title_short | Bacterial Immunogenicity Prediction by Machine Learning Methods |
| title_sort | bacterial immunogenicity prediction by machine learning methods |
| topic | protective immunogens machine learning immunogenicity prediction |
| url | https://www.mdpi.com/2076-393X/8/4/709 |
| work_keys_str_mv | AT ivandimitrov bacterialimmunogenicitypredictionbymachinelearningmethods AT nevenazaharieva bacterialimmunogenicitypredictionbymachinelearningmethods AT irinidoytchinova bacterialimmunogenicitypredictionbymachinelearningmethods |