| Publication No: | IN202211069259 [India] | Application No: | IN202211069259 |
| Title: | A process for selection and classification of drug targets from host pathogen protein-protein interaction data | ||
| Publication Date: | 31-05-2024 | File Date: | 30-11-2022 |
| Inventor(s): | Nirupma Singh; Sonika Bhatnagar | ||
| IPC Classification: | G06K 9/62, G16B 5/00, A61P 9/00, G16B 20/00, G16H 50/20 | ||
| Abstract: | Host-pathogen interactions play a crucial role in causing infections in humans. Many of these infections lead to adverse cardiac events. We have developed a random forest algorithm implemented for the classification of drug targets. The algorithm was trained using known hosts and pathogen drug and non-drug target proteins. We have standardized the method for collection and classification of protein physico-chemical, biochemical, structural, biological and functional parameters used for training the algorithm. In all, 68 pathogen and 73 host features were computed including network parameters. Further, data cleaning and normalization methods were used. After 10-fold cross-validation, the Random Forest classifier model achieved 99% accuracy with a ROC-AUC score of 0.99+/-0.01 for both pathogen and host. The Random Forest classifier was then used for selection of drug targets involved in Microbe Associated Cardiovascular Diseases. 331 host and 743 pathogen proteins were predicted as drug targets for Microbe Associated Cardiovascular Diseases. As drug target selection constitutes a starting point for drug development, this method has significant implications for the treatment of cardiac diseases caused by various microorganisms. Further, we have now implemented this algorithm for selecting drug targets for the cardiac effects of SARS-CoV-2 virus, an emerging complication of COVID-19 disease. This will help in development of novel drugs and repurposing of known drugs for COVID-19 associated heart disease. |
||