Re-Routing Drugs to Blood Brain Barrier: A Comprehensive Analysis of Machine Learning Approaches With Fingerprint Amalgamation and Data Balancing
Problem
Predicting whether existing drugs can cross the blood-brain barrier is difficult, hindering repurposing of drugs for neurological diseases.
Why it matters
Most drugs cannot cross the blood-brain barrier, so fast identification of permeable candidates could speed therapy for neurological conditions.
What the AI did
Machine learning models with mol2vec fingerprints were benchmarked on two datasets for blood-brain barrier permeability prediction, with SMOTE data balancing to improve specificity.
Results
AUC_ROC of 0.9453 (BBB dataset) and 0.9601 (B3DB dataset) for tree-based ensembles
Impact
Tree-based ML can accurately flag BBB-permeable drug candidates, supporting computational repurposing.
Record
| Venue | IEEE Access |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Mohammed Yusuf Ansari, Vaisali Chandrasekar, Ajay Vikram Singh, Sarada Prasad Dakua |
| Citations | 82 (OpenAlex, at import time) |
| References | 62 |
| Model or method | Random forest and extratrees with mol2vec fingerprints; SMOTE |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1109/access.2022.3233110 |
| Open access | gold |
| Keywords | drug repurposing, blood-brain barrier, machine learning, mol2vec, SMOTE, AUC |
Importance rating
real but narrower advance
Abstract states BBB permeability is a primary criterion for effective neurological therapy and aids drug repurposing; a useful screening advance, but no scale stated and results are benchmark-level.